S P A R C 2011Newsletter no36

January

STRATOSPHERIC PROCESSES AND THEIR ROLE IN CLIMATEA Project of the World Climate Research Programme

I C

Report on 2010 SOLARIS Activities and Future Plans

K. Matthes, Deutsches GeoForschungsZentrum Potsdam, and Freie Universität, Germany (matthes@gfz-potsdam.de) K. Kodera, Nagoya University, Japan (kodera@stelab.nagoya-u.ac.jp)

This report is aimed at describing SOLARIS activities in 2010, as well as outlining future plans. It is structured as follows: First, a summary of the SOLARIS workshop and the SCOSTEP side-meeting in 2010 will be given; second, open ques-tions that arose from those meetings will be discussed; and third, future studies will be presented. All SOLARIS activities and future plans are also available on the newly designed and regularly updated website: http://sparcsolaris.gfz-potsdam.de/.

Summary of the Second SOLARIS Workshop and SCOSTEP

Side-meeting in 2010

The second SOLARIS (SOLAR Influence for SPARC) workshop was held from 10-12 March 2010 and was hosted by the Ger-man Center for Geosciences in Potsdam, Germany. Approximately 38 participants from 8 countries participated in the two-and-a-half day meeting to review the latest results in the field of modelling the solar

influence on climate, and to decide about future coordinated SOLARIS modelling activities. Additionally, a SOLARIS side meeting was held during the SCOSTEP conference in July 2010 in Berlin, Germa-ny, with 23 participants, some of who had been unable to attend the spring meeting at GFZ Potsdam.

The first day of the March workshop in-

Group picture of the Second SOLARIS Meeting in March 2010 at the GFZ, Potsdam. Photo courtesy of Jan Dostal (GFZ, Potsdam).

Contents

Report on 2010 SOLARIS Activities and Future Plans by K. Matthes and K. Kodera ..............................................1

Report on the 7th SPARC Data Assimi-lation Workshop by S. Polavarapu and D. Jackson .............................................5

Report on WCRP Workshop on Seasonal to Multi-Decadal Predictability of Polar Climate by T. G. Shepherd et al..............11

Report on the SPARC DynVar Workshop 2 on Modelling the Dynamics and Vari-ability of the Stratosphere-Troposphere System by E. Manzini et al. .................19

The SPARC Data Initiative by M. Hegglin and S. Tegtmeier ...............22

Program of the Antarctic Syowa MST/IS Radar (PANSY) by K. Sato et al. .........23

Concordiasi: A Project Dedicated to the Polar Atmosphere by F. Rabier et al.....27

The High-Energy-Particle Precipi-tation in the Atmosphere (HEPPA) Model vs. Data Inter-comparison: Les-sons Learned and Future Prospects by B. Funke...............................................28

Stratospheric Change and its Role for Climate Prediction (SHARP): A Con-tribution to SPARC by U. Langematz et al. .....................................................32

Future Meetings...................................36

(January 11, 2011 / 08:55:48)

69967-1newsletter36-v2_New_p01.pdf .1

2

cluded a series of overview talks from in-vited speakers that were open to the gen-eral public. These overview talks covered a wide range of topics, including new esti-mates and reconstructions of solar variabil-ity in spectral and total solar irradiance go-ing back to 1610 (S. Solanki), preliminary results for the irradiance reconstruction over the last 11 000 years (Holocene), and it was mentioned that the sun is leaving its grand maximum in the current solar cycle 23. U. Cubasch presented modelling work with coupled middle atmosphere-ocean models for selected paleoclimate events. In particular, he showed results from the start and the end of the interglacial Eemian period as an example of the transition to an ice age, as well as results for the Holo-cene period, a climate optimum. The model simulations combined orbital forcings with solar variability to simulate the glacial and interglacial periods. K. Labitzke re-viewed and updated observational analy-sis of the solar signal in the stratosphere, focusing on the role of the QBO. Her first papers with van Loon in 1987 and 1988 are now on much firmer ground, with a gain of almost 60 years of data (QBO data have been reconstructed back to the mid-1940s). She highlighted the strong sum-mer signal. A. Brauer introduced a data base of annually laminated lake sediments at GFZ Potsdam, and showed solar signals from two selected records - the Meerfelder Maar Lake in the western part of Ger-many, and the Lake Ammersee in South-ern Germany - that confirm a sun-climate link that needs to be further investigated. B. Funke presented results from a SOLARIS-related project, the High-En-ergy Precipitating Particles in the Atmo-sphere (HEPPA) initiative (see also the HEPPA article in this issue). A first inter-comparison between different 2D and 3D models and MIPAS observations focused on the solar proton (SPE) event in 2003, the “Halloween” storm (Funke et al., 2010, in prep.). Since SPEs are sporadic events and likely do not have any long-term effects on the overall climate, new inter-comparisons will investigate the effect of energetic elec-tron precipitation (EEP) events, such as the EEP event during Northern Hemisphere winter 2009, which was a strong event in a dynamically active winter. EEP events are linked to geomagnetic activity and are modulated by the solar cycle.

L. Hood reviewed the origin of the tropical lower stratospheric ozone response to the solar cycle, currently a highly-discussed topic, since earlier 2D and 3D models have not been able to reproduce the observed vertical structure in the tropical solar signal in temperature and ozone. K. Kodera de-scribed some of the dynamical mechanisms through which small direct stratospheric ef-fects can indirectly affect the lower parts of the atmosphere down to the Earth’s sur-face, and proposed a mechanism for the modulation of the solar signal by the QBO. J. E. Kristjansson reviewed research on the impact of galactic cosmic rays on cli-mate, another proposed mechanism for how solar variability could influence cli-mate. It is now well established that elec-tric charge can enhance aerosol nucleation, and that nucleated aerosols may eventually grow to condensation nuclei. Recent stud-ies of Forbush decreases (rapid decreases in the observed galactic cosmic ray inten-sity following a coronal mass ejection) give different answers as to whether cosmic ray induced ionization influences clouds. There is no cosmic ray signal in aerosol nucleation events in Europe, little support from model studies, and global temperature is seemingly uninfluenced by cosmic rays. However, ongoing research at CERN may give new insights.

The following talks presented either ob-servational or modelling studies on the topic of solar influence. From the extended ERA-40 reanalysis, H. Lu confirmed the results of Labitzke and van Loon regard-ing the QBO modulation of the solar cycle. C. Blume investigated the dependence of the occurrence of stratospheric warmings in extended ERA-40 reanalysis on differ-ent natural variability factors (solar cycle, QBO, ENSO, volcanoes) with non-linear time series analysis, using an artificial neu-ral network. Y. Kuroda showed a solar cycle and QBO modulation of the southern annular mode (SAM).

A. Shapiro from the PMOD in Davos showed reconstructions of spectral and total solar irradiance variations from the Maunder Minimum to today, and came up with a much larger increase (6 W/m²) than other estimates, e.g. by J. Lean or S. So-lanki. S. Oberländer and A. Shapiro inves-tigated the effect of different spectral solar irradiance measurements on shortwave heating rates and circulation in stand-alone radiation code calculations and online cal-

culations with chemistry-climate models.

T. Reddmann talked about modelling of stratospheric chemistry during solar-in-duced NOx enhancements, observed with the MIPAS instrument onboard ENVISAT, in the Kasima model. Using a mechanistic model, I. Cnossen highlighted the impor-tance of gravity wave effects in modelling the solar signal propagation, as well as the importance of having a realistic representa-tion of stratospheric sudden warmings.

K. Matthes presented an overview about the SPARC CCMVal (2010) report, with special focus on the natural variability chapter and the comparison of the solar signal in the different observational and chemistry climate model analyses. It is still under discussion whether the vertical struc-ture of the tropical solar signal in ozone and temperature, especially the secondary maximum of the lower stratospheric signal, is related to non-linear effects, or aliasing of the solar cycle, the QBO, and ENSO.

A number of presentations of different chemistry-climate models focused on more detailed investigation of either the CCMVal reference simulations for the SPARC CCMVal report and the WMO ozone assessement (A. Kubin, Y. Yamashi-ta, G. Chiodo, K. Shibata), or idealized solar forcing experiments (C. Bell, S. Schi-manke), and new simulations for the next IPCC report (S. Watanabe). Results from coupled atmosphere-ocean models, includ-ing the middle atmosphere but not interac-tive chemistry, were presented by S. Mis-sios and T. Spangehl.

The SCOSTEP conference in July 2010 in Berlin provided an ideal forum for many of the SOLARIS participants to present their latest results. Some high-lights in-cluded the first presentation of solar influ-ence investigations in the new CMIP5 cou-pled atmosphere-ocean simulations from WACCM4 (D. Marsh), and a first com-parison of the new SOLARIS-proposed filtered experiments, which are similar to the CCMVal REF-B1 simulations but in-cluding non-correlated forcings from three CCMs (EMAC, WACCM, and MRI) (A. Kubin, K. Matthes, K. Shibata, K. Kodera, and U. Langematz).

Open Questions

Fruitful discussions and exchanges took place during the workshop and the side-

(January 11, 2011 / 08:55:51)

69967-1newsletter36-v2_New_p02.pdf .1

3

meeting. The following open questions emerged as prerequisites for a realistic rep-resentation of the solar signal:• What is the role of the QBO?• What is the role of the SSTs in the mean

climate (e.g., what effect do SSTs have on the occurrence of stratospheric warmings)?

• What is the role of a coupled ocean?• What is the impact of using different

spectral solar irradiance data sets?To approach these questions, new coordi-nated SOLARIS model studies were dis-cussed and are presented in the next sec-tion.

Future Studies

Proposed Coordinated Experiments

In order to approach the above questions and compare the response in different CCMs, we propose coordinated model studies designed to specifically investigate the solar cycle response. These studies will be in addition to the existing coordinated studies in the SPARC-CCMVal initiative, where natural and anthropogenic forcings have also been included to reproduce the recent past. The direct effect of the 11-year solar cycle in the upper stratosphere depends on a good representation of solar radiation processes in the radiative transfer and the photochemical parameterizations (e.g., Gray et al., 2010), and is reasonably simulated by the CCMs (SPARC CCMVal, 2010). However, indirect dynamical ef-fects in the tropical lower stratosphere and the extra-tropical stratosphere, as well as the extension of the signal into the tropo-sphere, are more difficult to reproduce and were not discussed in the CCMVal report.

The following coordinated studies are therefore proposed (further details, such as input data and a detailed ex-perimental description, can be found at http://sparcsolaris.gfz-potsdam.de/):

A) Coordinated Model Runs to Investigate Aliasing of Different Factors in the Tropi-

cal Lower Stratosphere

Recently, the discrepancy between mod-elling studies and observations regarding the vertical structure in the tropical solar signal, as shown by the WMO (2007), has been reduced in both CCMVal-1 refer-ence simulations (Austin et al., 2008) and CCMVal-2 simulations (SPARC CCMVal, 2010, Chapter 8). Similarly, other recent

simulations with CCMs reproduce the observed vertical structure in the tropical stratosphere, but only with a (prescribed) QBO, time-varying solar cycle conditions and constant SSTs (Matthes et al., 2007; Matthes et al., 2010), or in a CCM with fixed solar cycle conditions, with or with-out an internally-generated QBO (Schmidt et al., 2010). It is still unclear why a verti-cal structure in the solar signal appears, and whether it is related to non-linear interac-tions or arises from contamination by other signals (QBO, tropical SSTs).

To eliminate possible aliasing between the solar cycle and the QBO, as well as be-tween the solar cycle and the SSTs, and/or the QBO and the SSTs, the REF-B1 CCMVal experiments were repeated with filtered SST and/or QBO data. The QBO signal (2-3 years) and solar cycle signals (larger than 10 years) have been filtered out of the SST data set used as a lower bound-ary for the REF-B1 simulations. Similarly, the QBO data were filtered to retain only periods between 9 - 48 months and exclude signals related to ENSO or the solar cycle.

Currently, two CCMs with a prescribed QBO (EMAC, WACCM), and one with internally generated QBO (MRI) have fin-ished one ensemble of the modified REF-B1 experiments.

B) Coordinated Model Runs to Study the Uncertainty in Solar Forcing

Uncertainties in the solar irradiance could have a large impact on the simulation of the climate system. The solar irradiance data compiled by J. Lean (Lean, 2000) are

most frequently used for model simula-tions. However, in the wavelength range important for ozone chemistry (200 - 400 nm), there are differences of up to 20% to the estimate of Krivova et al. (2006). Fur-ther uncertainties arise from new measure-ments from the SIM instrument onboard the SORCE satellite, which shows a com-pletely different spectral distribution than expected, with possible implications for solar heating and ozone chemistry (e.g., Haigh et al., 2010).

The proposed coordinated CCM experi-ments would include:1. A control (time slice) experiment with

either Lean (standard) or SIM solar ir-radiance data,

2. Idealized experiments with enhanced solar UV forcing in certain spectral ranges, i.e., an increase of 5% between 200 and 300 nm, and an increase of 1% between 300 and 400 nm.

An example of possible implications on solar heating rates from an increase of 1% in solar irradiance data from Lean be-tween 300 and 400 nm is shown in Figure 1. The annual mean difference between a 7-year control run and increased solar ir-radiance experiments with WACCM shows an impact on the shortwave heating rates of 0.05 K, and on subsequent variables (not shown). These differences are one third of those reported from solar cycle studies and are therefore not negligible.

It is important to investigate the reliability of current SOLARIS irradiance data rec-ommendations for CMIP5 and CCMVal,

Figure 1: Annual mean short wave heating rate differences between a 7-year control run and a 7-year enhanced irradiance experiment. Note that the values have been divided by 10 for better readability (courtesy of G. Chiodo, University of Madrid).

(January 11, 2011 / 08:55:53)

69967-1newsletter36-v2_New_p03.pdf .1

4

and test the sensitivity of different radia-tion and photochemistry models to differ-ent spectral irradiance data sets.

CMIP5 Recommendations and Analysis

SOLARIS provided recommendations for solar irradiance data sets to be used for the CMIP5 simulations. In particular, recom-mendations were made for the pre-indus-trial control runs, as well as the future pro-jections for models that either change total solar irradiances (TSI) only, and climate models that prescribe spectrally resolved solar irradiance variations. Since a num-ber of long, coordinated simulations with high-top CCMs exist within CMIP5, we are planning an inter-comparison of these runs with respect to the solar signal. The tropospheric solar signal in the high-top CMIP5 simulations will be investigated in order to gain insight into typical patterns, e.g. the tropical Pacific response (Meehl et al., 2009) and sensitivity on model charac-teristics.

Goals

SOLARIS provides an excellent plat-form for the coordination and discus-sion of solar-related studies, as well as to provide recommendations for the solar irradiance data used to drive middle at-mosphere and climate models within the SPARC CCMVal initiative and the CMIP5 simulations. The data portal will be extend-ed and maintained, especially with meth-ods for the inclusion of energetic particles. It will foster and initiate detailed studies on the solar UV and TSI mechanisms, as well as high energy particle events (in collaboration with HEPPA). Commu-nication and collaboration between the SOLARIS middle atmosphere community and the climate community has begun, and will be intensified and extended, especially in light of the fact that a number of climate models that include the middle atmosphere are participating in the next IPCC scenario simulations, in support of the next IPCC report in 2013.

Acknowledgments

This article is dedicated to the memory of Christopher Bell, who had a promising ca-reer ahead of him and was unfortunately the victim of a tragic accident in June 2010.

We would like to thank WCRP/SPARC for their support, as well as for the sponsor-ship from the Helmholtz Association in the frame of the Helmholtz-University Young Investigators Group of KM. Also, we ap-preciate the logistical help from Veronika Söllner from GFZ Potsdam.

References

Austin, J., et al., 2008: Coupled chemistry climate model simulations of the solar cy-cle in ozone and temperature, J. Geophys. Res., 113, doi:10.1029/2007JD009391.

Gray, L. J., J. Beer, M. Geller, J. D. Haigh, M. Lockwood, K. Matthes, U. Cubasch, D. Fleitmann, G. Harrison, L. Hood, J. Luterbacher, G. A. Meehl, D. Shindell, B. van Geel and W. White, 2010: Solar in-fluences on climate, Rev. Geophys., 48, doi:10.1029/2009RG000282.

Haigh, J. D., A. R. Winning, R. Toumi and J. W. Harder, 2010: An influence of solar spectral variations on radiative forc-ing of climate, Nature, 467, 696-699, doi:10.1038/nature09426.

Krivova, N. N., S. K. Solanki and L. Floyd, 2006: Reconstructions of solar UV irradi-ance in cycle 23, Astron. Astrophys., 425, 631-639.

Lean, J., 2000: Evolution of the Sun’s spec-tral irradiance since the Maunder mini-mum, Geophys. Res. Lett., 27, 2425-2428.

Matthes, K., D. R. Marsh, R. R. Garcia, D. E. Kinnison, F. Sassi and S. Walters, 2010: Role of the QBO in modulating the influ-ence of the 11 year solar cycle on the atmo-sphere using constant forcings, J. Geophys. Res., 115, doi:10.1029/2009JD013020.

Matthes, K., K. Kodera, L. Gray, J. Aus-tin, A. Kubin, U. Langematz, D. Marsh, J. McCormack, K. Shibata and D. Shindell, 2007: Report on the first SOLARIS work-shop: 4–6 October 2006, Boulder, Colo-rado, USA, SPARC Newsletter 28, 19–22.

Meehl, G. A., J. M. Arblaster, K. Matthes, F. Sassi and H. van Loon, 2009: Amplify-ing the Pacific climate system response to a small 11-year solar cycle forcing, Science, 325, DOI:10.1126/science.1172872.

Schmidt, H., G. P. Brasseur and M. A. Giorgetta, 2010: Solar cycle signal in a general circulation and chemistry mod-

el with internally generated quasibien-nial oscillation, J. Geophys. Res., 115, doi:10.1029/2009JD012542.

SPARC CCMVal, 2010: SPARC CCMVal Report on the Evaluation of Chemistry-Climate Models, V. Eyring, T. G. Shepherd, D. W. Waugh (Eds.), SPARC Report No. 5, WCRP-X, WMO/TD-No. X, http://www.atmosp.physics.utoronto.ca/SPARC.

World Meteorological Organization, 2007: Scientific assessement of ozone depletion: 2006, Global Ozone Res. and Monit. Proj. Rep. 50, Geneva, Switzerland.

(January 11, 2011 / 08:55:57)

69967-1newsletter36-v2_New_p04.pdf .1

5

Report on the 7th SPARC Data Assimilation Workshop

21-23 June 2010, Exeter, UK

S. Polavarapu, Environment Canada, Canada (Saroja.Polavarapu@ec.gc.ca)D. Jackson, Met Office, UK (david.jackson@metoffice.gov.uk)

The Seventh Stratospheric Processes And their Role in Climate (SPARC) Data As-similation (SPARC-DA7) workshop held in Exeter, England during 21-23 June 2010 was the latest in a series of annual meet-ings that bring together data assimilators, users of assimilation products, and experts in modelling, measurements and process studies. Data assimilation requires know- ledge of measurement and model errors, which, in turn, require knowledge of the true underlying system – the middle atmo-sphere. Therefore, advancement of assimi-lation techniques requires interaction of as-similators with dynamicists, chemists and users of assimilation products. One of the ways in which we broaden the participation in these meetings beyond data assimilators is by inviting experts (who usually are not data assimilators) to present lectures and to promote discussion along certain themes. This year the themes were: “seamless pre-diction”, stratosphere-troposphere coupling and tropospheric constituent assimilation.

“Seamless prediction” refers to the goal of improving predictions on weather to climate time scales. It is also a focus area and common goal of both the World Weather Research Programme (WWRP) and the World Climate Research Pro-gramme (WCRP). One aspect of seamless prediction is the use of data assimilation for understanding model errors. The sec-ond theme was motivated by the fact that at operational weather centres, the impor-tance of the stratosphere is primarily seen through its impact on the troposphere. Fi-nally, there is increasing demand for air quality and environmental forecasting and assimilation. Thus, the last theme offered an opportunity to better connect the strato-spheric chemical assimilation community with the tropospheric air quality assimila-tion community.

Seamless Prediction and Model Error

In addition to evaluating climate models statistically, it is also useful to consider

their predictions on short time scales. A definition of seamless model assessment is the use of a wide variety of time scales to assess and improve the representation of processes within a model. The basis for the notion of seamless assessment lies in the fact that model errors on a given time scale can first appear on shorter time scales. For example, K. Williams showed that cloud regime properties are local but tend to appear at the first time step and persist for about 5 days, and that errors in the Madden-Julian Oscillation (MJO) can appear by day 5. For this reason, the Transpose-AMIP project (supported by the WGCM and WGNE of WCRP and WWRP, respectively and chaired by Williams) has as its goal the evaluation of climate models on short time scales. ECMWF reanalyses for 2008-9 provide the initial conditions and 64 5-day forecasts are run. By assess-ing model parameterizations against spe-cific events, climate models participating in CMIP5 can be assessed in terms of their ability to depict fast processes. The short-term predictive ability of climate models can also be compared.

P. Telford discussed “nudging”, or relax-ation, of a CCM version of the Met Office UM to ECMWF analyses. This approach is another very effective way of identifying model biases, in this case near the tropo-pause. Furthermore, this approach is useful for diagnostic studies, for example in the attribution of ozone changes in the after-math of the Mt Pinatubo eruption. Future applications might include local “nudging” to examine how reducing errors in one lo-cation (e.g., European blocking) might af-fect model errors elsewhere.

Data assimilation can be used to infer as-pects of the dynamical or chemical behav-iour of the atmosphere, though it is always important to assess the limitations of such an approach. Y. Orsolini presented a tech-nique based on data assimilation to esti-mate chemical polar ozone loss in the 06/07 Arctic winter. By comparing the results with another ozone loss estimate based on

CTM simulations, he was able to show the strengths of the data assimilation approach, but also highlight issues with transport in the fields produced by data assimilation, which could largely be addressed in future by repeating the experiments using 4D-Var rather than 3D-Var.

R. Menard examined different methods of background error covariance calcula-tion applied to chemical data assimilation. He found that the NMC method (National Meteorological Centre) works in a denser observation network and under advection dynamics only, while the forecast differ-ence approach (sometimes know as the Ca-nadian Quick method) can easily introduce smaller-scale variances and correlations. For the Desroziers method, the estimation of the error variances is sensitive to the mis-specification of the observation error correlation scale (much more than to mis-specification of the corresponding quantity for the background error). The representa-tion of background errors was also a key part of the talk by T. Milewski, who is developing an ensemble Kalman filter for stratospheric chemical data assimilation. Using synthetic temperature and ozone observations, he showed that there were some benefits in the localization of the background error covariances, although this comes at the expense of a loss of mass/wind balance.

Of relevance to Milewski’s work, but of potential importance to all areas of ensem-ble data assimilation, was the presentation by C. Bishop, in which the concept of the “outer loop” currently used for 4D-Var was derived and interpreted in the ensemble context. The outer loop in ensemble data assimilation provides an opportunity to re-run the ensemble about an ensemble mean that is closer to the truth. This might be at-tractive in areas of strong gradients, such as near fronts, which may be poorly rep-resented by any of the ensemble members. Experiments with a simple soliton model showed how re-running the ensemble in the outer loop allows the ensemble covariance

(January 11, 2011 / 08:55:58)

69967-1newsletter36-v2_New_p05.pdf .1

6to better approximate the covariance of the distribution of historical forecasts given the truth.

If the general idea behind seamless predic-tion is the use of various time scales to as-sess and minimise model error, then data assimilation also has a role to play in seam-less assessment of models. Data assimila-tion can be used to estimate uncertain pa-rameters in parameterization schemes (such as gravity wave drag schemes) and thereby reduce model error. Alternatively, at ECMWF, model error is now being directly estimated along with initial conditions. Y. Tremolet noted that a constant model error term has been estimated operationally since September 2009. The estimated model er-ror is largest in the stratosphere and ap-pears to be systematic, rather than random. The addition of model error estimation to the assimilation system reduces the size of analysis increments, but the estimate does not always reflect model error. In one ex-ample, bias in aircraft observations over Denver was picked up in the model error term. In another example, deficiencies of balance operators in background error co-variances were found to be similar to the model error estimate. Figure 1 shows the monthly mean analysis increment with (centre panel) and without (left panel) a model error term. Increments are smaller with the model error term. However, the

right panel shows that simply removing the balance operators from error covariances produces similar mean analysis increments to that obtained with a model error term. Thus, perhaps with the amount of observa-tions assimilated, balance is being captured in the troposphere and there is little need for additional balance constraints. In the stratosphere, the covariance estimates are noisy and the balance operators may need improving. Future work involves allowing the model error term to be time varying, over long windows, effectively approxi-mating a Kalman smoother. A key aspect of future improvements in NWP forecasts will rely on the accurate characterization of model error, just as it is with climate mod-els.

New applications of data assimilation to problems of long time scales were present-ed by L. Neef and K. Miyazaki. L. Neef described the problem of assimilating Earth orientation parameters to constrain climate models since wind and mass changes influ-ence the Earth’s angular momentum which affects its wobble and rotation rate. The challenge is using a globally integrated quantity to update state variables. K. Mi-yazaki described a new way to estimate carbon fluxes at the Earth’s surface using a state-augmented ensemble Kalman filter. In his scheme, meteorological variables were estimated simultaneously with CO2

and CO2 fluxes, improving CO2 estimates. A variety of data sources were assimilated (from surface flask measurements, aircraft and GOSAT) making this perhaps the most advanced carbon flux estimation system in the world.

Observations and Impacts

An important role for data assimilation in climate science lies in assessing the impact of proposed Earth observations (most frequently satellite missions). As pointed out by W. Lahoz, the quantity of interest is the incremental benefit of a proposed new observation set over the existing global Earth observation network. To assess such a benefit, Observing System Simulation Experiments (OSSEs) are typically performed, and are increasingly being required of new missions early in the design phase. In OSSEs, synthetic measurements from the proposed instrument are assimilated and compared with a control cycle without the new measurements. In both cycles, the most realistic observation network (i.e., most of the existing currently available observations) is used. While the results of OSSEs are not always clear because they are model dependent and because the issue of biases (from models or observations) are usually not addressed, OSSEs can still be valuable. For example, OSSEs for SWIFT (Stratospheric Wind

0.5

MetgraF-30-20-15-10-7-5-3-2.0-1.5-1.0-0.7-0.4-0.2-0.10.10.20.40.71.01.52.0357101520

90N 90S

-1.5

-0.7

-0.7

-0.7

-0.2

-0.2

-0.2

-0.2

-0.2

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.4

0.4

Pre

ssur

e (m

b)

90N 90S

-1.5

-0.7

-0.7

-0.2

-0.2

-0.2

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.40.4

0.40.4

0.4

1.0

100070050030020015010070503020

10

5.0

2.0

1.0

0.2

0.1

0.05

0.02

0.01

90N 90S

-3.0

-1.5

-1.5

-0.7

-0.7

-0.2

-0.2

-0.2

-0.2

-0.2

-0.2

-0.2

-0.2

-0.2

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.1

0.10.10.10.1

0.4 0.4

0.4

0.4

0.4

0.4

0.40.4

1.0

1.0

1.0

2.0

1

2

345678910

1112131415161718202224262830323436404460

Figure 1: Monthly mean analysis increment with ECMWF weak constraint. Mean analysis increments indicate bias since random errors should average to zero. ERA-Interim (left panel) are compared to those obtained with a weak constraint (centre panel) and those obtained by removing balance operators applied to background error covariances (right panel). (Figure courtesy of Y. Tremolet, ECMWF.)

(January 11, 2011 / 08:56:02)

69967-1newsletter36-v2_New_p06.pdf .1

7

Interferometer for Transport Studies) found a significant benefit of SWIFT winds in the tropical stratosphere (Lahoz et al., 2005). Lahoz also noted that OSSEs can be used to compare the potential impact of two proposed instruments,

showing that MAGEAQ (Monitoring the Atmosphere from Geostationary orbit for European Air Quality) would be more beneficial than MTG/IRS at observing lower tropospheric CO and O3. Using an ensemble data assimilation approach, H.

Körnich found the impact of ADM winds on zonal wind forecasts to be comparable to that of radiosondes in the Arctic, the tropics and over oceans (compare top and bottom panels of Figure 2). However, the impact of using different vertical sampling strategies (targeting the UTLS or the stratosphere) was unclear. Interestingly, vertical propagation of information is seen in that improved forecasts in the troposphere at day 4 lead to improved stratospheric forecasts at day 7, with this time scale being consistent with that of large scale Rossby waves (Figure 2c, top and bottom).

The assimilation of new satellite instruments requires the preparation of new forward models relating measured variables to model (or analysed) variables. For operational assimilation, common software (such as fast radiative transfer models) is often used by several forecasting centres. H. Lewis presented a new software package to pre-process GPS radio occultation data from levels 1a, 1b and 2 aimed at operational centres, but such software would be equally valuable for research assimilation systems. Detailed information may be found at http://www.grassaf.org. In addition to providing software, bending angles, refractivity, temperature, pressure and humidity profiles are available in near-real time with additional products available off-line.

When satellite retrievals are assimilated, as is often the case with observations of con-stituents, there is danger of assimilating a priori content instead of the information from the measurement. To avoid this, A. Kaiser-Weiss described a new method of assimilating retrievals in which an inter-face step is added between the retrievals and their assimilation in 4-D variational schemes. In this step, the information of observations is isolated and measurement space is transformed so errors are uncorre-lated. This new method will be tested at ECMWF with real measurements.

Stratosphere-Troposphere coupling

As operational weather forecasting centres have increased model lid heights into the mesosphere, in recent years, stratospheric data assimilation has become the purview of operational centres. Since operational centres are mainly concerned with the prediction of weather in the troposphere,

Radiosonde - impacta) b)

c) d)

Global u impact Tropical u impact

NH-extratropics u impact SH-extratropics u impact

-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04

Forecast time (hours)

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Forecast time (hours)

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

1/2 ADM-impacta) b)

c) d)

Global u impact Tropical u impact

NH-extratropics u impact SH-extratropics u impact

Forecast time (hours)

Pre

ssur

e (P

a)

1

3

10

30

100

300

850

-0.02 -0.015 -0.01 -0.005 0 0.005 0.01 0.015 0.02

6 12 24 48 72 24021619216814412096Forecast time (hours)

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Pre

ssur

e (P

a)

1

3

10

30

100

300

8506 12 24 48 72 24021619216814412096

Figure 2: Zonal wind impact as estimated from the reduction of the ensemble forecast spread due to removal of radiosonde data (top 4 panels:a,b,c,d) or the addition of ADM-Aeolus data using the strato-vertical sampling (bottom 4 panels:a,b,c,d). Only half of the expected ADM-data amount was included. The area averages include only significant spread changes on a 95%-confidence level. Though the vertical sampling scenario fo-cuses on the stratosphere, it also contains the troposphere, but with less dense sampling than the tropospheric sampling scenarios (Figure courtesy of H. Körnich, Stockholm U.)

(January 11, 2011 / 08:56:14)

69967-1newsletter36-v2_New_p07.pdf .1

8

stratospheric data assimilation is primarily valued for its ability to impact tropospheric forecast quality. The fact that stratospheric dynamics can influence the troposphere on monthly to seasonal to decadal time scales was noted by A. Scaife in an overview presentation. Stratospheric events, such as sudden warmings, volcanic eruptions or ozone depletion, and signals such as the quasi-biennial oscillation can influence the extra-tropical NAO signal. Furthermore, the stratosphere can modulate tropospheric signals such as ENSO, again influencing the NAO. Thus the stratosphere effectively has a “memory” which can be exploited to enhance seasonal to decadal predictions. Nevertheless, not all models running sea-sonal or decadal predictions adequately resolve the stratosphere. Thus, CliVar’s WGSIP (Working Group on Seasonal and Interannual Prediction) has added a new component to its international Coupled Historical Seasonal Forecasting Project (CHFP) in which the additional benefit of resolving the stratosphere is assessed in the context of seasonal forecasts. For decadal prediction, CMIP5 will have some models that resolve the stratosphere.

An overview of mechanisms that could explain the influence of the stratosphere on the troposphere was presented by A. Charlton-Perez. Figure 3 (from Reichler et al., 2005) illustrates the various stages involved in producing a downward propagating signal from the stratosphere, such as those seen in Baldwin and Dunkerton (2001)’s dripping paint

figure. In stage 1, a tropospheric wave is launched near the surface. It propagates up to the stratosphere in stage 2, increasing in amplitude as density decreases. The wave dissipates, interacting with the mean flow in the stratosphere in stage 3, leading to downward propagating wind anomalies in stage 4. Stage 5 describes the stratosphere-troposphere coupling stage, for which various mechanisms have been proposed. Charlton-Perez noted that while the mechanisms to explain stage 5 are still in dispute, none of the proposed mechanisms can be ruled out. He also wondered whether it is important to know which mechanism (e.g., direct influence of stratosphere PV anomalies on the troposphere, stratospheric zonal wind shear causing reflection of planetary waves back into the troposphere, modulation by stratospheric zonal winds of baroclinic wave life cycles or eddies) explains stratosphere-troposphere coupling or whether it is sufficient to know that they all can occur.

While stratosphere-troposphere coupling is usually examined on time scales of a few weeks or longer, the influence of the stratosphere or the tropospheric medium range weather forecasts (up to 10 days) was considered by S. Polavarapu using the Canadian Meteorological Centre’s operational forecasting system. In com-paring the forecast skill of a low top system (with model lid height of 10 hPa) to a high top system (with a model lid height of 0.1 hPa), a clear downward propagation of forecast skill from day 1 to day 10 was seen in extra-tropical regions. The high-top system resulted in a dramatic 75% reduction of geopotential forecast error standard deviation (measured against radiosondes) in the lower stratosphere, with 5-10% reduction in the troposphere. Most of the improvement (in the troposphere and the stratosphere) was due to the model changes, not to the addition of extra observations in the upper stratosphere (AMSU ch. 11-14, GPS RO from 30-40 km). This result is consistent with the notion that information can propagate upward in a model that is capable of simulating realistic stratospheric dynamics. The impact of the stratosphere on tropospheric forecasts in the 5-10 day range was also seen by S. Mahmood when assessing the skill of the Met Office model with a 40 km top to one with an 80 km top. However, the additional benefit of increased stratospheric resolution on tropospheric forecast skill was modest,

and more apparent on the 10-15 day range. Thus, once the stratosphere is reasonably simulated, further improvements may result in smaller incremental benefits on tropospheric forecasts.

N. Zagar presented an analysis of Kelvin waves in various atmospheric analyses. The normal modes method she used to perform the analysis has the advantage of being able to test impacts of changes in model physics and data assimilation on vertical energy propagation and wave properties. Here, she showed that there are large differences in the climatologies of Kelvin waves calculated from NCEP, ECMWF and other analyses, which are related to unreliable tropical circulations in these analyses. Focusing on ECMWF analyses for 2007-2009, she found evidence for stratospheric Kelvin waves, which propagate energy upward with periods about 16 days. Horizontal phase speeds are between 20 m/s and 30 m/s. Maximal Kelvin wave activity is around the tropopause region in the Pacific.

D. Jackson examined low ozone events (LOEs) in the southern summer strato-sphere seen in ozone analyses based on EOS MLS data. Particularly interesting were the deep LOEs seen when tongues of low ozone at 31 and 100 hPa were super-posed. The superposition is probably due to the tilt of baroclinic medium scale (wave-numbers 4-6) waves present between the two levels. Transport means that the LOEs are not simply focused over the pole, but tend to be seen preferentially over the Wed-dell Sea. The reason for this is unclear.

Stratosphere-Mesosphere coupling

Vertical coupling of the troposphere, strato-sphere and mesosphere occurs through ver-tically propagating gravity waves. Liu et al. (2009) showed that whether perturba-tions were introduced to the lower atmo-sphere or to the mesosphere, the resulting error growth appearing after 2 days was similar. Thus, errors in the lower tropo-sphere are connected to mesospheric er-rors through vertically propagating grav-ity waves. Since gravity waves reduce predictability decay (Ngan et al., 2009) in rotating stratified turbulence, K. Ngan asked whether the increasing role of grav-ity waves in mesospheric energy spectra also leads to increased predictability (or reduced predictability decay). Using the

0°20°W5

10

25

50

75

100

150

200

500

stratosphere

troposphere

z

t

1

2

3

4

5t0 t0+∆t t0+τ

Figure 3: Schematic illustration of Tropo-sphere-Stratosphere-Troposphere events. (1) Forced pulse of planetary waves occur-ring over time Δt; (2) upward-propagating waves; (3) dissipation and breaking of waves; (4) induced downward-propagating anomalies; and (5) tropospheric response at time lag τ > Δt. (From Reichler et al., 2005.)

(January 11, 2011 / 08:56:16)

69967-1newsletter36-v2_New_p08.pdf .1

9

Met Office unified model with a lid at 60 km, Ngan showed the relative kinetic ener-gy spectra in the troposphere, stratosphere and mesosphere. Predictability decay was found to be modestly slower in the me-sosphere than in the troposphere, but nu-merical issues of whether gravity waves are properly resolved were raised.

Vertical coupling of the middle atmosphere is also illustrated by stratospheric sudden warming events, which are driven by plan-etary waves propagating up from tropo-sphere. Manney et al. (2008; 2009) have shown that such events can sometimes re-sult in large changes in stratospheric extent with the stratopause reaching heights of 75 km or more. Not surprisingly, data assimi-lation systems with model lids or sponge layers below 75 km have difficulty simu-lating such events. In an effort to improve the GMAO model’s ability to fully simu-late stratopause evolution, L. Coy and S. Pawson raised the GMAO model lid from 0.01 hPa or 80 km to roughly 100 km. The Fomichev non-LTE radiation scheme was added, background error covariances were extended to the mesopause, and other ad-justments were made (such as to gravity wave drag parameters). With the assimi-lation of MLS temperatures, the extended GMAO system was able to better describe the January 2006 stratospheric sudden

warming event. The same event was also used by S. Ren to assess the addition of SABER temperatures to the CMAM as-similation system. Though the CMAM had been able to capture the stratopause height changes noted by Manney (2008) without assimilating mesospheric measurements (Polavarapu et al., 2008, Fig. 4), Ren no-ticed that when mesospheric temperatures from SABER are assimilated, the simula-tions are less sensitive to the presence or absence of a nonorogaphic gravity wave drag scheme. In addition, M. Keller noted that without the assimilation of mesospher-ic temperatures, local instabilities such as the mesospheric 2-day wave are not cap-tured. Keller further considered how the assimilation of mesospheric measurements is able to capture the 2-day wave, wheth-er through direct sampling of the wave at 6-hourly intervals, or through adjustment of the zonal mean flow. While evidence pointed to the latter explanation in January 2007, a mystery remains in that the Feb-ruary 2007 2-day wave was absent from CMAM analyses even when SABER tem-peratures were assimilated.

In addition to GMAO, the Met Office is also studying the impact of raising their model lid (from 70 to 80 km). D. Long showed that changes to the non-orographic gravity wave drag scheme (to parameters,

to make it momentum conserving) and the conversion of the turbu-lent dissipation of grav-ity waves to a heating resulted in changes to zonal mean biases of the model. Since weather forecasting model do-mains may now reach the mesopause, their domains can overlap with upper atmospheric physical models used for space weather. A. Bushell noted that space weather products and services are now being developed in Eu-rope, and that the mete-orological data assimi-lation community has a potential role to play in assessing upper atmo-sphere physical models, and perhaps in coupling the lower atmosphere to

upper atmospheric models. As a first step in this direction, Bushell used SABER and MLS temperatures to assess Met Office and CMAT2 (an upper atmosphere model) fields in the region where their domains overlap (100 hPa to 80 km). Preliminary conclusions were that satellite measure-ments such as those from SABER and MLS were very valuable for assessing middle at-mospheric and lower thermospheric model performance.

Constituents

D. Jones discussed the assimilation of tro-pospheric ozone (and precursors) in an air quality prediction system. Assimilating tro-pospheric ozone and CO (from TES) was generally effective in reducing the ozone bias in the analyses, but it is also a good method for pinpointing model biases. The assimilation significantly increased ozone and reduced the bias in the free troposphere, which they attributed to an under-estimate of lightning NOx emissions, based on the CO analysis. Figure 4 (see colour plate I) shows the impact of assimilating TES data on surface ozone over the USA. Compari-son with independent observations shows that after assimilation, the bias in surface ozone was reduced in the western USA, but surface ozone was still overestimated in the eastern USA. This was attributed to an overestimate of surface emissions of NOx in the model. Improving the representa-tion of these emissions improved the ozone analyses (Figure 5 - colour plate I). The speaker raised an interesting question from this result: should we use data assimilation for parameter estimation (as here), or just as a means of reducing forecast errors (as in NWP), or both? F. Baier’s talk focused on lower stratospheric ozone and showed the benefits of assimilating ozonesonde data, at least near the sonde stations. Improved analysis may be achieved via further tuning of both observation and background error covariances.

The assimilation of multiple species in Jones’ talk certainly highlights the benefit of complementary information, since with-out CO assimilation, it would have been hard to attribute model bias to the errone-ous NOx emissions. In a similar manner, Q. Errera aimed to assimilate stratospheric ozone to constrain inorganic Cl in the stratosphere. While it shows that the con-straint of assimilated ozone on modelled HCl is weak, the constraint was found to

0°20°W

50°N

60°N

5

10

25

50

75

100

150

200

500

Figure 6: The locations where AIRS data (black dots) and IASI data (blue dots) have been rejected due to excessive aerosol contamination during the 12z 4D-Var assimilation window on 15th April. Superimposed is a 39-hour MACC forecast from 14th April (verifying with the centre of the same 4D-Var assimi-lation window) predicting the trajectory of simulated aerosol injection at 5 km. The units in the legend indicate only a relative magnitude of column aerosol tracer. (Figure courtesy of Tony McNally, Reima Eresmaa and Johannes Flemming, ECMWF.)

(January 11, 2011 / 08:56:18)

69967-1newsletter36-v2_New_p09.pdf .1

10

be sufficient to allow the analyses of HCl to reproduce the observed HCl trend in the upper stratosphere lower mesosphere dur-ing the three years period of assimilation.

In the area of aerosol assimilation, Y. Prad-han and K. Ngan described the develop-ment of a dust assimilation scheme for the Met Office Southern Asian LAM. SEVIRI aerosol optical depths are assimilated. Ini-tial trials using 3D-Var and run for Jan / Feb 2010, show promising results. Future work will extend the data assimilation to 4D-Var and run more extensive trials (e.g., for major dust events).

A. Benedetti gave an overview of aero-sol assimilation work at ECMWF (based on MODIS aerosol optical depths) with a focus on the analysis of the volcanic ash cloud from the recent Eyjafjallajökul erup-tion. Non-operational simulations with a simple representation of the source of the ash showed reasonable agreement against AIRS and IASI aerosol data. Most of the uncertainty in these plume forecasts is caused by the assumptions about the injec-tion height. Further work to include aerosol optical depths from SEVIRI and AATSR is promising but biases in these products have to be identified and corrected. Operational analyses of the Eyjafjallajökul plume suf-fered problems because quality control re-jected most of the aerosol observations in the plume (see Figure 6) and because there is no way of representing the source of the volcanic plume. In readiness for future eruptions, a future operational system may include the option to switch to a “volcanic eruption mode”, which would incorporate relaxed aerosol observation quality control and a volcano location / plume height da-tabase.

Discussion

The SPARC data assimilation working group members are continuing to work in the areas of interest to SPARC: vertical coupling of the stratosphere with the tro-posphere and the mesosphere, constituent assimilation and observing system design. New applications are continuing to emerge, such as environmental prediction (e.g., aerosol and dust forecasting), operational carbon flux estimation, and even extracting long time scale information about climate from Earth rotation observations. Issues that continue to be raised include the reli-ability of analyses in the tropics, and how

to utilise assimilation techniques to im-prove model error. While meteorological data assimilation may be primarily used to obtain initial conditions for weather fore-casting, it may be that for air quality or cli-mate applications, data assimilation is most valuable for obtaining model parameters including constituent sources and sinks or model errors. This use of data assimila-tion comes under the “seamless prediction” banner and is of interest to both the WCRP and the WWRP. As the WCRP and SPARC evolve into their post-2013 form, an im-portant question for the present is how will the data assimilation working group evolve and what role will it play?

Next meeting

The next SPARC Data assimilation work-shop will be held in Brussels, Belgium dur-ing 20-22 June 2011. The local organizer is Quentin Errera. Themes and invited speakers will be announced in the next few months.

References

Lahoz, W. A., R. Brugge, D. R. Jackson, S. Migliorini, R. Swinbank, D. Lary and A. Lee, 2005: An observing system simulation experiment to evaluate the scientific merit of wind and ozone measurements from the future SWIFT instrument. Q. J. Royal Me-teor. Soc., 131, 503-523.

Liu, H.-L., F. Sassi and R. R. Garcia, 2009: Error growth in a whole atmosphere Cli-mate model. J. Atmos. Sci., 66, 173-186.

Manney, G. L., M. J. Schwartz, K. Krüger, M. L. Santee, S. Pawson, N. J. Lee, W. H. Daffer, R. A. Fuller, and N. J. Livesey, 2009: Aura Microwave Limb Sounder ob-servations of dynamics and transport dur-ing the record-breaking 2009 Arctic strato-spheric major warming. Geophys. Res. Lett., 36, DOI:10.1029/2009GL038586

Manney, G. L., K. Krüger, S. Pawson, K. Minschwaner, M. J. Schwartz, W. H. Daf-fer, N. J. Livesey, M. G. Mlynczak, E. E. Remsberg, J. M. Russell III and J. W. Waters, 2008: The evolution of the stra-topause during the 2006 major warming: Satellite data and assimilated meteorologi-cal analyses. J. Geophys. Res., 113, DOI: 10.1029/2007JD009097

Martin, R. V., C. E. Sioris, K. Chance, T. B. Ryerson, T. H. Bertram, P. J. Wooldridge,

R. C. Cohen, J. A. Neuman, A. Swanson and F. M. Flocke, 2006: Evaluation of space-based constraints on global nitro-gen oxide emissions with regional aircraft measurements over and downwind of east-ern North America, J. Geophys. Res., 111, doi:10.1029/2005JD006680.

Ngan, K., P. Bartello and D. N. Straub, 2009: Predictability of rotating stratified turbulence. J. Atmos. Sci., 66,1384-1400.

Parrington, M., D. B. A. Jones, K. W. Bow-man, A. M. Thompson, D. W. Tarasick, J. Merrill, S. J. Oltmans, T. Leblanc, J. C. Witte and D. B. Millet, 2009: Impact of the assimilation of ozone from the Tropospher-ic Emission Spectrometer on surface ozone across North America, Geophys. Res. Lett., 36, doi:10.1029/2008GL036935.

Polavarapu, S., E. Farahani, G. Manney, N. McFarlane and T. G. Shepherd, 2008: Re-port on the joint SPARC workshop on data assimilation and International Polar Year. SPARC Newsletter no. 30, 27-33.

Reichler, T., P. J. Kushner and L. M. Polvani, 2005: The coupled stratosphere–troposphere response to impulsive forcing from the troposphere. J. Atmos. Sci., 62, 3337–3352, doi:10.1175/JAS3527.1

(January 11, 2011 / 08:56:20)

69967-1newsletter36-v2_New_p10.pdf .1

11

Report on WCRP Workshop on Seasonal to Multi-Decadal Predictability of Polar Climate

25-29 October 2010, Bergen, Norway

T. G. Shepherd, University of Toronto, Canada (tgs@atmosp.physics.utoronto.ca)J. M. Arblaster, Centre for Australian Weather and Climate Research, Australia (jma@ucar.edu)C. M. Bitz, University of Washington, USA (bitz@atmos.washington.edu)T. Furevik, University of Bergen, Norway (tore@gfi.uib.no)H. Goosse, Université Catholique de Louvain, Belgium (hgs@astr.ucl.ac.be)V. M. Kattsov, Voeikov Main Geophysical Observatory, Russia (kattsov@main.mgo.rssi.ru)J. Marshall, Massachusetts Institute of Technology, USA (jmarsh@mit.edu)V. Ryabinin, WMO, Switzerland (VRyabinin@wmo.int)J. E. Walsh, University of Alaska Fairbanks, USA (jwalsh@iarc.uaf.edu)

Background and purpose of workshop

Over the last few decades, the polar re-gions have exhibited some of the most striking changes in the observed climate record. Whilst the Arctic has warmed, as expected from the polar amplification of greenhouse-gas (GHG) induced warming arising from the ice-albedo feedback, the observed rate of summer-time sea-ice re-treat in the Arctic is at the upper limit of climate model predictions. At the same time, Antarctic sea-ice extent is observed to be increasing, contrary to the model pre-dictions. The clearest observed changes in the Antarctic, which are associated with the poleward shift and intensification of the summer-time mid-latitude jet, are primar-ily attributed to the ozone hole, which im-plies that the observed trends will weaken substantially or could even reverse in the coming decades as the ozone hole recov-ers. However, natural variability in polar regions is large, with substantial power at multi-decadal time scales, and manifests it-self in large-scale “modes” whose physical nature and causality are not clear. There are even suggestions of an inter-hemispheric “see-saw”. As a result, it is difficult to de-termine how much of the recent behaviour might be due to natural variability.

The observed and predicted changes in the polar regions have significant implications. In the Arctic, sea-ice changes will directly impact shipping, resource extraction, pol-lution, and coastal erosion, affecting the lives of inhabitants and those who oper-

ate in that region. In the Antarctic, ocean circulation changes will affect the rate of carbon uptake in the Southern Ocean, and could have implications for the stability of the West Antarctic ice shelf. There is, therefore, a pressing societal need to im-prove the reliability of climate model pre-dictions in polar regions, including both the response to anthropogenic forcings and our understanding of the decadal time scale variability. The spatial-temporal coherence of this variability offers the hope that some component of it might actually be predict-able, given knowledge of the initial state, and the initial state could also be important for the response of the polar regions to an-thropogenic forcing. However, at this point we do not really know which measure-ments are most important for constrain-ing that state. In addition to the societal benefits that would result from improved predictions, we would also be in a better position to explain the variability in the evolving climate record.

Because of the strong coupling that exists in the polar regions between ocean, sea ice, troposphere and stratosphere, it is neces-sary for all these scientific communities to work together in order to make significant progress on these problems. This was the motivation behind the WCRP Workshop on Seasonal to Multi-Decadal Predictability of Polar Climate, which brought together approximately 80 experts on polar climate variability and predictability from around the world, representing not only the above-mentioned range of physical disciplines but also observations, theory, processes, and

modelling, and with a bi-polar, global per-spective. The purpose of the workshop was to summarize the current state of knowl-edge and identify concrete steps to improve our predictive capability in polar regions1. The workshop was hosted by the Bjerknes Centre at the University of Bergen, and was formally opened by the Rector of the University, Sigmund Grønmo. JSC Chair Tony Busalacchi also provided welcoming remarks on behalf of the WCRP.

Reports on scientific sessions

A preliminary session provided some back-ground context. T. Shepherd presented the scientific motivation for the workshop (as described above), and emphasised the role of the oceans, sea ice, land surface and stratosphere as inherently stable parts of the climate system, with significant in-ertia, which provide non-trivial boundary conditions (with memory) for the variabil-ity that is ultimately driven by the unstable troposphere (Figure 1). This suggests that the key to improved prediction is under-standing and accurately representing the sources of memory within the different climate system components, and the feed-backs between them. The stratosphere is a special case because it also represents a source of external forcing (ozone deple-tion, ozone changes due to solar variabil-ity, aerosols due to volcanic eruptions, and possibly geoengineering), in addition to 1The workshop programme and copies of the presentations can be found at http://www.atmosp.physics.utoronto.ca/SPARC/PolarWorkshop/presentations_bergen.htm

(January 11, 2011 / 08:56:22)

69967-1newsletter36-v2_New_p11.pdf .1

12

GHG forcing. B. Kirtman reviewed recent progress in seasonal to decadal prediction, which lies between the two extremes of weather prediction (dependent entirely on initial conditions, and essentially deter-ministic) and centennial time scale climate projections (dependent mainly on exter-nal forcings). Modern seasonal prediction relies almost entirely on the predictability of ENSO, and forecast skill is generally confined to temperate latitudes. There are believed to be untapped sources of seasonal predictability in the stratosphere, in sea ice, and in the land-surface (including snow cover), which are all operative at high lati-tudes — stratosphere-troposphere coupling is strongest in polar regions — so inclusion of these processes should improve predic-tive skill at higher latitudes. On decadal time scales, the extra-tropical ocean is also believed to represent an untapped source of predictability. G. Boer highlighted the rap-idly growing scientific interest in decadal predictability, though noted that the decadal time scale was more of a human than a physical time scale. He reviewed recent studies of predictability in polar regions from a “potential predictability” perspective, which uses a “perfect model” framework to identify what fraction of the year-to-year changes might consist of long-time scale processes (including the forced component) that are potentially predictable given sufficient knowledge of the initial state, as opposed to unpredictable climate noise. These studies hint at some potential predictability in polar regions (Figure 2),

but it is not yet clear how large or useful this will be. Some open issues raised in the talks by Kirtman and Boer included the use of a multi-model ensemble (which seems invariably to outperform the “best” models, even for the same ensemble size), how to best combine dynamical and statistical ap-proaches, and how to objectively define the potentially predictable, as opposed to noisy part of the signal. In the end, predictability is a property not just of the physical system but also of the filter we apply to it, which depends on the application.

Session 1 was devoted to the mechanisms that rule sea ice variability, the way they are represented in models, and the processes

that may help us in providing useful pre-dictions. H. Goosse discussed the observed and simulated variations over the last cen-turies. He insisted that the last 30 years are not necessarily representative of the full range of variability of the system and thus collecting and analysing longer time series is needed, in particular to adequately evalu-ate the variability simulated by models. R. Kwok presented a comparison of simulated and observed ice motion and ice transport. He highlighted that many models have strong biases that need to be reduced in or-der to improve our ability to make good pre-dictions. C. Bitz discussed different mecha-nisms that could lead to predictability in the

Figure 1: Evidence for the role of the stratosphere in modulating tropospheric teleconnections. A 5-member ensemble of AMIP-type simulations with the Météo-France model (shading, with the thick blue curve the ensemble mean and the dashed lines +/- 1 standard deviation) is not able to reproduce observed (black curve) interannual variations in the DJF surface Northern Hemisphere Annular Mode (NAM) over the 1971-2000 time period, represented here by the principal component of surface pressure north of 20°N, when constrained only by SSTs (left panel), but does so extremely well when the extra-tropical stratosphere is nudged to ERA-40 reanalyses (right panel). R is the ensemble mean anomaly correlation coefficient with ERA-40. From Douville (2009, GRL).

Figure 2: Evidence for decadal predictability of surface temperature at high latitudes from low-frequency internal variability, based on a “perfect model” diagnosis. The shad-ing shows the fraction of the internally generated temperature variance accounted for by decadal and longer time scales within a multi-model ensemble of unforced control simulations in the CMIP3 archive. The presumption is that these long time scales are “potentially predictable” with sufficient information and knowledge (see also Boer and Lambert, 2008, GRL).

(January 11, 2011 / 08:56:28)

69967-1newsletter36-v2_New_p12.pdf .1

13

system, analysing ensembles of simulations with a coupled general circulation model (GCM) and observations. Re-emergence of anomalies in different seasons related to sea surface temperatures (SSTs) (linked with ice concentration changes) or changes in ice thickness appeared to be particularly prom-ising for predictions of Northern Hemi-sphere sea-ice area on time scales of a few months (Figure 3). J. Overland proposed a hypothesis in which a reduced summer ice cover would lead to a warmer fall in the Arctic, inducing a decrease in atmospheric geopotential height and a large-scale reor-ganization of the atmospheric circulation that may be characterized by a low index of the Arctic Oscillation. If this mechanism is valid, it would have a strong impact on predictions at the seasonal scale, as well as on long term changes, because of the strong decreasing trend in Arctic ice extent pro-jected for the next decades.

M. Raphael described sea-ice variability in the Southern Ocean and its links with atmospheric changes. Sea ice is influenced

by all the known modes of atmospheric variability in the Southern Hemisphere: the Southern Hemisphere Annular Mode (SAM), the Pacific Southern America (PSA) pattern, the Semi-Annual Oscillation (SAO), and the Zonal Wave 3 (ZW3). How-ever, none of them explains a great deal of the sea ice variance integrated over the whole Southern Ocean, probably because those modes were defined as atmospheric modes, rather than in terms of their impact on sea ice. F. Massonnet discussed the im-portance of model physics, resolution and forcing in simulations of Arctic and Ant-arctic sea-ice variability performed with an ocean-sea ice model driven by atmospheric reanalyses. A good forcing was found to be essential. Model physics appeared crucial in order to reproduce well the variability in the Arctic, in particular in summer, while the improvements brought by a more sophisti-cated model were less clear in the Southern Ocean. In the range of resolutions tested (between 0.5 and 2°), the resolution of the sea ice model was not the most critical issue.

K. Giles discussed the avail-able satellite observations of sea-ice thickness and how they could be used to make skillful decadal predictions. Analysing observed sea-ice thickness and concentration, she showed that the summer extent is well correlated with the thickness of the follow-ing winter, but not with that of the previous winter, sug-gesting that summer extent could help in estimating the next winter’s ice thickness.

Session 2 began with a re-view by D. Holland of chal-lenges in understanding and modelling cryospheric pro-cesses, with an emphasis on ocean ice-shelf interactions relevant to, for example, Greenland glacial fjords and the West Antarctic ice sheet. He emphasised the key role of intrusions of warm deep water in ice-shelf melt. Un-fortunately, warm deep wa-ter can neither be seen from space nor inferred from gravity. K. Heywood then discussed the physics and observations of Antarctic Bottom Water formation,

and the extent to which global climate models are able to capture the large-scale circulation features in the Southern Ocean. She argued that while climate models seem to provide a reasonable representation of the transport of intermediate water, they are much worse at representing the transport of surface and bottom water. Moreover, cli-mate models form deep water incorrectly through open-ocean convection.

The next two talks reviewed open ocean processes and the dynamics of the Antarc-tic Circumpolar Current (ACC). S. Gille discussed observed recent changes in the hydrographic structure (including ocean heat content) and frontal positions of the ACC (Figure 4), emphasising that we do not really understand the causality under-lying those changes. While studies with non-eddy-permitting ocean models sug-gest that the shifts in the ACC front have been driven by the changes in the SAM, others have argued that ocean eddies buf-fer this effect and lead to a very different sensitivity. This is clearly an important is-sue to resolve in the future. J. Marshall discussed the central role of the Southern Ocean in the upwelling branch of the glob-al overturning circulation. It is important to understand which parts of this circulation are relaxational (i.e., can accommodate changes elsewhere) and which are ‘choke points’ that require forcing. In particular, it is not entirely clear the extent to which atmospheric wind stresses and heat fluxes, and winter sea-ice cover, may respond to as well as drive Southern Ocean upwelling.

The focus then shifted to the Northern Hemisphere. A. Proshutinsky pointed to the role of the wind driven Beaufort Gyre, alternating between a strong anticyclonic regime accumulating ice and fresh water, and a weaker cyclonic regime releasing the fresh water to the North Atlantic and influencing the overturning circulation. C. Mauritzen emphasised the revolution that has occurred in recent years in near real time data acquisition in the Arctic and Sub-Arctic, with more than 30,000 Argo, Ice-Tethered Profiler (ITP), glider, and seal-borne CTD profiles obtained since 2001. The new data will give us the op-portunity to narrow down the uncertainties in ocean heat content, fresh water content, and density both for reanalysis and opera-tional products. S. Østerhus reviewed the direct measurements of mass, heat and salt exchanges between the North Atlantic and

Jan May Sep Jan May Sep Jan May Sep0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8Standard Deviation of sea ice area 106 km2

Figure 3: Evidence for seasonal predictability of sum-mer-time sea ice based on “perfect model” experiments. The black, gray, blue and light blue curves in the figure show the growth of the standard deviation of Northern Hemisphere sea-ice area across an ensemble of model simulations where each member was initialized with identical sea-ice, ocean, and land surface conditions, starting from different points in the year. The black dashed curve is the saturation level of the standard de-viation from a long control run. The solid curves lie be-low the black dashed curve, indicating that sea ice area is potentially predictable for up to a year in advance. The initial loss of potential predictability is faster for start dates in the summer season. Nonetheless, based on these experiments, perfect knowledge of the initial conditions in winter only offers modest predictability of sea ice area in the following summer. Figure courtesy of Cecilia M. Bitz, University of Washington.

(January 11, 2011 / 08:56:41)

69967-1newsletter36-v2_New_p13.pdf .1

14

the Arctic. More than 10 years of measure-ments show no trends in volume transports, but there has been a rapid increase in heat

and salt fluxes.

T. Eldevik noted that the exchange of mass, heat and salt over the Greenland-Scotland ridge can be de-scribed mathematically by three forced conservation equations. With this ap-proach, the sensitivity of the transports to changes in hydrography or forcing can be tested. Climate models that fail to reproduce the three distinct water masses at the ridge will respond differently to forcing per-turbations. While oceanic responses to atmospheric forcing are well docu-mented in observations and models, mechanisms for oceanic forcing of the atmosphere outside the

tropics are less well understood. A. Czaja proposed a new mechanism for ocean-

atmosphere coupling in the extra-tropics. While the textbook version is that a warmer ocean surface heats the atmosphere above and creates a baroclinic response, surface temperatures can set the atmospheric lapse rate over the warm western boundary cur-rents and thus communicate the tempera-ture signal throughout the entire tropo-sphere. These moist neutrality situations are currently not parameterized in global atmospheric models, leaving a potential for prediction yet to be fully investigated.

Session 3 examined the role of the strato-sphere in predictability. P. Kushner dis-cussed mechanisms for coupling between the stratosphere and the troposphere. He distinguished between direct effects of stratospheric changes on the troposphere, and indirect effects whereby the state of the stratosphere affects teleconnections (e.g., ENSO) within the troposphere. He also emphasised the impact of model bi-ases, e.g., models with too-long annular-mode time scales exhibit overly strong annular-mode responses to external forc-

Circumpolar meanºL

atitu

de

SAZ/STZSAFPFSACCF

1992 1994 1996 1998 2000 2002 2004 2006 2008

0.6

0.5

0.4

0.3

0.2

0.1

0

-0.1

-0.2

-0.3

-0.4

Figure 4: Evidence for decadal predictability in the South-ern Ocean from long-term (presumably forced) changes. The different curves show estimates of changes in the mean latitudinal position of the fronts that comprise the ACC, as inferred from satellite altimeter measurements. (Here, SAZ/STZ is the Sub-Antarctic Zone/Sub-Tropical Zone; SAF is the Sub-Antarctic Front; PF is the Polar Front; and SACCF is the Southern ACC Front). The ACC has shifted poleward by about 60 km over the last 15 years. From So-kolov and Rintoul (2009, JGR).

Figure 5: Evidence for decadal predictability of changes in Southern Hemisphere summer-time atmospheric circulation from ozone depletion and GHG changes. 30-year trends calculated from reconstructions of the 20th century Southern Hemisphere Annular Mode (SAM) index (left panels) show that the recent summer-time trends are unprecedented in the historical record, indicating that they are a response to external forcings. The CMIP3 model simulations (right panels) suggest that the dominant component of the forcing in this season is due to stratospheric ozone depletion. In other seasons, ozone has a negligible impact and the recent trends are just becoming significant in the fall season, but not in the spring, where the simulated trends are too strong. No significant winter trends are evident in reconstructions or simulations (not shown). The dotted lines represent the range of internal climate variability from the model’s pre-industrial control simulations (left panels), rescaled by the square root of six (the number of non-ozone models) (right panels). From Fogt et al. (2009, J. Clim.).

(January 11, 2011 / 08:56:43)

69967-1newsletter36-v2_New_p14.pdf .1

15

ings. S. Yoden reviewed insights obtained from mechanistic models concerning in-ternal and externally forced variability of the winter-time stratospheric polar vortex, which dominates stratospheric variability. The use of a mechanistic model permits very long simulations (e.g., 15,000 years), which are needed to fully characterize the PDFs of stratospheric variability because of the large amount of decadal variability and high degree of intermittency in Strato-spheric Sudden Warmings (SSWs). T. Shepherd reviewed basic aspects of strato-spheric variability, and summarised the various physical mechanisms for memory in the stratosphere on both seasonal and interannual time scales, including tropi-cal-polar coupling. Stratospheric models (whether simple or complex) and observa-tions both exhibit strong decadal time scale variability, but it has yet to be determined how predictable this variability is.

J. Perlwitz discussed the impact of the Antarctic ozone hole on the Southern Hemisphere (SH) high-latitude summer-time troposphere, which is by far the clear-est instance of a stratospheric influence on surface climate, and is a predictable signal since it is associated with stratospheric halogen loading. The surface impact of the ozone hole involves a strengthening and poleward shift of the tropospheric jet (rep-resented by the SH Annular Mode (SAM)), and has implications for the ocean circula-tion, which are beginning to be studied. The anticipated recovery of stratospheric ozone over the coming decades implies that this component of the recent climate trends will be reversed, with the latest model studies suggesting a near cancellation for summer-time trends between the effects of ozone recovery and GHG forcing over the next half century. J. Arblaster discussed the re-sponse of the SAM, which controls much of SH climate, to future GHG and ozone forc-ing, emphasising that the response to ozone forcing is mostly confined to the summer season. She found that the circulation re-sponse to GHG forcing is strongly related to climate sensitivity and arises more from the warming of the tropical upper tropo-sphere, which previous studies have shown induces dynamical (momentum flux) feed-backs through a strengthened subtropical jet, rather than from polar cooling.

M. Sigmond addressed the question of whether the ozone hole might explain the observed increase of Antarctic sea-ice ex-

tent. In his coupled model simulations, with a non-eddy-permitting ocean model, the ozone hole led, instead, to a reduced sea ice extent. This decrease is consistent with a mechanism involving enhanced off-shore Ekman sea ice transport arising from the stronger westerlies. A poster by C. Bitz also found a decrease in sea-ice extent in response to the ozone hole employing a dif-ferent model. However, she found that the response was significantly smaller when the ocean model resolution changed from non-eddy-permitting to eddy-permitting, owing to a significant reduction in the poleward heat transport response at higher horizontal resolution. This result is consistent with the ‘buffering’ effect of eddies that was dis-cussed by S. Gille (see above). J. Jones presented a reconstruction of the SAM in-dex over the entire 20th century. This re-cord is important because there is a paucity of long data sets for SH high latitudes. She found that in DJF the recent increase of the SAM index was unprecedented in the his-torical record, so presumably a response to forcing (which models suggest has mainly come from the ozone hole), whereas in MAM the recent increase was large but still within the range of natural variability (Figure 5). No SAM trends were identified in either JJA or SON.

Session 4 focused on predictability of the Arctic climate system, and covered ocean-

atmosphere exchanges, mid-latitude-Arctic coupling, high-latitude-terrestrial predict-ability, sensitivities and feedbacks in the Arctic system, and the use of models for prediction. X. Zhang showed how an ob-servational constraint (the sensitivity of Arctic sea-ice extent to air temperature) could be used to narrow the range of future sea-ice projections obtained from global climate models. Zhang also showed how the recent loss of summer Arctic sea ice is part of a broader Rapid Change Event in-volving a shift of the atmospheric circula-tion. H. Tanaka discussed the role of the Arctic Oscillation (AO), which explained about half the Arctic warming from the 1960s to the 1990s. He showed that the AO is almost dynamically orthogonal to the “global warming” component of the re-cent Arctic change.

M. Karcher documented the variability of the Atlantic Water inflows and outflows, for which the Arctic Ocean acts as a switch-yard (Figure 6 - colour plate II). While these inflows have subpolar origins, the Nordic Seas impose their imprint. Karcher showed that North Atlantic inflow anoma-lies may impact the deep water overflows about 10-15 years after entering the Arctic Ocean, implying some potential predict-ability of overflow variability. K. Shi-mada showed that the recent reduction of sea ice in the western Arctic Ocean is

2.0

1.5

1.0

0.5

0.01 2 3 4 5 6 7

r=-0.55w/o outlier: r=-0.8

2000 2020 2040 2060 2080 2100 2120

LambdaArea (0-0.5)

0.8

0.6

0.4

0.2

0.0

Exp

. Var

. in

SS

IE

Year

( )

War

min

g (K

)

winter inversion strength (Tdiff - 850 hPa and surface)

Figure 8: Mechanisms controlling spread in the Arctic climate change predictions of the CMIP3 models. Left: Relationship between winter inversion strength and annual-mean Arctic warming by the 22nd century (A1B emissions scenario). The stronger the inver-sion, the more heat is lost by cooling to space (mainly from clear-sky conditions), and the smaller the overall annual-mean warming. The observed inversion strength lies at the left end of the model range, suggesting the models may have unrealistically high levels of negative long-wave feedback. Right: Fraction of explained variance in Arctic sea-ice extent changes in CMIP3 models from present day to the indicated year, from the radiative feedback parameter λ (mainly related to inversion strength) (black) and the climatologi-cal extent of thin sea ice (grey). These two parameters, which can both be constrained by observations, account for nearly all the predicted changes in sea-ice extent, with the latter dominating in the first few decades and the former dominating later in the century. From Boé et al. (2009, J. Clim.; 2010, Clim. Change Lett.).

(January 11, 2011 / 08:56:45)

69967-1newsletter36-v2_New_p15.pdf .1

16

due to a combination of three factors: heat, preconditioning, and winds. The inflow-ing Pacific Summer Water is a source of heat, but its impact on sea ice is amplified by wind-driven changes in the Beaufort Gyre dynamics and the interplay with re-duced ice concentrations, Ekman pumping, and topography. M. Serreze discussed the broader issue of polar amplification, and showed that its strongest manifestation in the cold season and the marginal ice zone is indicative of a contribution of a feedback arising from the reduced sea ice. There is a need for coordinated model experiments to assess the impacts of the reduced sea-ice extent on the atmospheric circulation else-where in the Northern Hemisphere.

In the first of the “terrestrial” presentations, D. Lawrence showed that in CCSM3 21st century A1B simulations, the rate of warm-ing over Arctic land areas is enhanced by 1-2°C/decade in autumn and winter during periods of rapid sea-ice loss (Figure 7 - colour plate II). He showed that the same seasonality and spatial pattern of Arctic

land warming was also found in AMIP-type experiments forced by the sea-ice loss projected by CCSM3 by the end of the 21st century. Lawrence also showed that 21st century simulations are accompanied by substantial changes in permafrost, as defined by the ground temperature at 3 m depth. A. Slater addressed the simulation of snow and soil temperature within a data assimilation framework. There are severe problems with the available data coverage for these variables, especially in the case of snow water equivalent and depth, despite the potential importance of these variables for predictability on seasonal time scales. Finally, H. Morrison showed that climate models poorly simulate Arctic clouds, es-pecially the partitioning of condensate into the liquid and ice phases. A key question is whether the frequency of occurrence of dif-ferent cloud states can be related to certain parameters available at the grid-cell scale. Large-Eddy-Simulation experiments can be exploited for this purpose.

A. Hall used a suite of CMIP3 simulations to assess the predictable component of Arctic change that is anthropogenically driven. Polar amplification is seen in the surface air tem-perature, but not in the heat content of the upper ocean, pointing to atmospheric processes as the key to the large spread in the models’ polar amplification. The main predictor of a model’s response to GHG forcing is the longwave feedback parameter under clear-sky conditions. The models’ low-level stratification is closely tied to this feedback, and the models with strong near-surface stratification show relatively little warm-ing because of strong long-wave cooling. In general, the models’ surface-based inversions are too strong. Hall further showed that the spread in models’ rates of sea-ice loss is related to two factors: the initial area of thin (<0.5 m) sea ice, and the longwave feedback parameter (Figure 8). J. Christensen raised the is-sue of potential nonlinearity

in the systematic errors of regional climate models, identifying cases where systematic errors in surface temperature had a strong dependence on temperature over particular European regions. Further work is needed to place these dependences into a frame-work of GHG-induced warming. A. Rinke reported on an ensemble of 15-month hind-cast simulations starting in March and Sep-tember of various years of the 1979-2009 period. The experiments included various combinations of atmospheric, sea ice/SST, and snow initializations. One of the key findings was that certain atmospheric con-ditions are more predictable than others.

Session 5 consisted of two parts: one on seasonal predictability involving sea ice or snow as predictors, and one on seasonal to decadal predictability involving fully cou-pled global climate models. M. Baldwin reviewed the evidence for the influence of stratospheric winter-time variability on surface weather regimes. Weak and strong stratospheric vortex events have been shown to influence the frequency distribu-tion of AO/NAO conditions up to 60 days after the events. Unfortunately the seasonal forecast models all under-estimate strato-spheric variability, indicating that there is still a way to go before the maximum forecast skill from stratospheric effects is realised. Y. Orsolini used the coupled ECWMF seasonal prediction system to show that sea-ice anomalies provide some predictability of Arctic surface air tempera-ture during autumn and early winter (Fig-ure 9), consistent with Lawrence’s infer-ences (see above) using GCM studies, and also that autumn sea ice variability can in-duce deep temperature anomalies through-out the troposphere and circulation changes influencing East Asia early winter climate. Orsolini also used an atmospheric GCM to show that Eurasian autumn snow cover can influence atmospheric wave trains over the North Pacific and eventually over the North Atlantic. J. Cohen pointed to au-tumn snow cover over Eurasia as a precur-sor for stratospheric variability. More snow and a strengthened Siberian High appear to strengthen the Rossby-wave flux into the stratosphere, weakening the polar vortex and thus favouring negative AO situations. Cohen also presented an experimental pre-diction of the winter 2010/2011 conditions based on his approach.

In parallel with such observational and reanalysis studies, efforts are being made

0 20 40 60 80

Forecast day

2-da

y-m

ean

AC

C

1.0

0.5

0.0

-0.5

Anomaly Correlation Coefficient

Figure 9: Evidence for enhanced seasonal predictability of Arctic surface air temperature during fall and early winter 2007 from prescribing sea-ice extent in ensemble hind-casts based on the ECMWF coupled seasonal prediction system. Record low summer-time sea-ice extent occurred in 2007. The 2-day mean anomaly correlation coefficients are shown as a function of forecast day (starting October 1, 2007), and calculated over high latitudes (60°N-90°N). Each black vertical bar is the envelope of the 5-member hindcasts using the prescribed 2007 sea ice, while the five grey vertical bars on its left are the envelopes of the 5-member hindcasts using prescribed, but “erroneous” or scrambled, sea-ice extent from the five preceding years (2002 to 2006). From Orsolini, Senan, Benestad and Mel-som, to be submitted.

(January 11, 2011 / 08:56:47)

69967-1newsletter36-v2_New_p16.pdf .1

17to improve the ocean/ice assimilation and forecasting systems. Two different meth-ods were presented; one using the Ensem-ble Kalman Filter approach (F. Counillon), and a second using a four-dimensional variational approach (F. Kauker). Using data assimilation, it is possible to identify key parameters or areas that are particu-larly sensitive to perturbations, and thus guide process studies or measurement campaigns. In recent years there has been a large increase in studies related to decadal climate forecasts. Three talks were given on this subject, all demonstrating that hindcast experiments do show promising skill both in real and idealised experiments. J. Jung-claus focused on the sources and impacts of variations in the Atlantic meridional overturning circulation. Time scales and mechanisms differ between models, and more work is needed to identify those that work in the real world. D. Smith showed that the North Atlantic also plays an ac-tive role through Subpolar Gyre dynamics, with links to tropical convection and hurri-cane frequency. Of particular interest is the near collapse of the Subpolar Gyre around 1995, seen from altimetry and downstream hydrography, which was discussed by E. Hawkins. The fact that climate mod-els reproduce this event in forecast mode (Figure 10) indicates that the anomalous atmospheric forcing that year (record low

NAO after many pos-itive years) played a smaller role than pre-viously believed. J. Walsh gave the final talk in the session. He showed that model ensembles are gener-ally more skillful in reproducing Arctic climate variations than single models, but only to a certain extent. The skill of the ensembles is re-duced if the models with the largest biases are included. In gen-eral the models with the best performance also tend to show a stronger sensitiv-ity to GHG forcing. Some predictability is expected from low-frequency variability in the Arctic climate.

Synthesis

We understand many of the physical sourc-es of predictability in the polar climate sys-tem. For sea ice, memory resides in sea-ice thickness, rather than sea-ice extent, and spring-time ice-thickness anomalies can re-emerge in the fall with summer-time mem-ory provided by the ocean. The initial sea-ice thickness distribution is the main control on modelled Arctic sea-ice loss for the first half of the 21st century. For snow, memory resides in snow depth (or snow water equiv-alent). There is longer-term memory in per-mafrost, whose disappearance can lead to an albedo feedback through rapid growth of shrubs. For the ocean, SST anomalies have a seasonal memory while longer-term memory resides in the heat and salinity of subsurface water masses, which provide a mechanism for lagged teleconnections. In the Antarctic there is also substantial mem-ory provided by the baroclinic component of the ACC, which exerts a control on the Atlantic Meridional Overturning Circula-tion. For the atmosphere, there is seasonal memory in the stratosphere, which modu-lates tropospheric variability, because of the long radiative time scales in the lower stratosphere and the strong seasonal cycle of stratospheric polar variability. There is also longer-term memory in tropical strato-

spheric winds, manifested in part by the QBO. The Antarctic ozone hole has been the principal driver of summer-time trends in SH high-latitude surface climate over the last few decades, which may cease or even reverse as the ozone hole recovers over the coming decades.

Although the field is still in its infancy, early results concerning the extent of polar predictability show promise. Operational seasonal prediction systems for the Arctic show the impact of summer-time sea ice and fall Eurasian snow-cover anomalies, and September Arctic sea-ice extent ap-pears to be predictable given knowledge of the spring-time ice thickness or early to mid-summer sea-ice extent. Stratospheric Sudden Warmings provide further predict-ability during winter and spring once they occur, although the extent to which they are themselves predictable is still unclear. On longer time scales, studies of potential predictability within a “perfect model” framework suggest multi-year predictabil-ity of the internal variability over the high-latitude oceans in both hemispheres, and the first attempts at decadal prediction have identified the Atlantic subpolar gyre as a key source of predictability, with a telecon-nection to tropical Atlantic SSTs.

What we lack is a good understanding of many of the feedbacks between the differ-ent components of the climate system. The precise dynamical mechanisms of strato-sphere-troposphere coupling remain to be elucidated, although they appear to be well represented in models that have sufficiently good climatological mean states. The ro-bust surface responses to stratospheric vari-ability and trends are in surface winds and mean-sea-level pressure gradients, which are dynamically controlled; the surface tem-perature responses, which are more ther-modynamically controlled, are less clear except where they result from advection. The response of Arctic sea ice to surface winds appears to be well understood, but the origin of the overly pole-centric Beau-fort Gyre in climate models, which induces significant biases in sea-ice export through the Fram Strait, is not well understood.

Although the basic mechanisms of Arctic amplification of GHG-induced warming, which involve feedbacks from sea ice and ocean, are well understood, there are large uncertainties in the magnitude of the sur-face temperature response arising from un-

Figure 10: Evidence for decadal predictability of North Atlan-tic upper ocean heat content based on the successful predic-tion of the North Atlantic subpolar gyre warming in 1995/1996 with the DePreSys ensemble prediction system. Upper ocean (0-500 m) temperature anomalies averaged over the North At-lantic subpolar gyre for a DePreSys hindcast started from June 1995 initial conditions is shown in blue, which successfully cap-tures the warming in the observations (black, taken from the Met Office ocean analysis (Smith and Murphy, 2007)). All tempera-ture anomalies are relative to a 1941-1996 climatology. From PhD thesis of Jon Robson, University of Reading, 2010.

(January 11, 2011 / 08:56:49)

69967-1newsletter36-v2_New_p17.pdf .1

18

certainties in the response of Arctic clouds and systematic model biases in boundary-layer stability. While global ocean models generally have a good representation of intermediate-water transport, they have a very poor representation of the transport of surface and bottom water, and incorrectly form deep water by open-ocean convec-tion. This could compromise the realism of the response of the ocean circulation to sur-face buoyancy forcing. In the SH, the re-sponse of the ACC and of poleward ocean heat transport to surface wind trends seems to be very different in eddy-permitting and non-eddy-permitting ocean models, sug-gesting that the latter may have non-con-servative eddy parameterizations that do not correctly “buffer” the ocean response to wind stress forcing.

As a result of all these weaknesses in our knowledge, we do not well understand the physical causality of the large-scale modes of polar variability that are evident in the observed record. This compromises our ability to design appropriate observation, assimilation, and modelling systems for polar prediction, and to explain the ob-served record.

Unfortunately, we lack many of the key observations needed to constrain the pre-sumed sources of polar predictability; ex-amples include snow depth and snow water equivalent (estimates from passive micro-wave instruments are widely regarded as useless), sea-ice thickness (estimates from laser altimetry may be acceptable, but they are not currently available in real time), polar ocean state estimates including Ant-arctic warm deep water, and stratospheric tropical winds. An exception is the salin-ity and heat anomalies entering and exiting the Nordic (GIN) Seas, which appear to be well constrained by hydrographic data in the limited number of communicating channels. However, there has been an ex-plosion of new subsurface ocean observa-tions in the last decade from Argo floats and from the recent “seal” network, which are revolutionising the polar ocean observ-ing system. These observations provide new opportunities for model validation — probably best performed in observation rather than model space, to avoid introduc-ing errors from interpolation, and with a focus on “process-oriented” diagnostics that are not overly sensitive to the time pe-riod considered — and offer the potential for vastly improved estimates of the ocean

state, a prerequisite for polar predictability. Nevertheless, since the “repeat cycle” for seasonal and especially decadal predic-tions is rather long, prediction systems will continue to be tested in hindcast mode, for which our poor historical knowledge of the ocean and sea-ice initial states will surely represent a major limitation.

Next steps

In considering what can be done to make progress in polar predictability, it needs to be kept in mind that it is not the job of WCRP to coordinate climate science. Nor is there much point in making unsolicited research recommendations. Rather, the WCRP aims to identity those aspects of climate science that benefit from interna-tional coordination. That means identify-ing particular gaps, typically where efforts by individual scientists or groups have run into a wall because of the lack of a wider effort. Since the WCRP has no staff of re-searchers, high-impact initiatives address-ing those gaps need to be developed that can rally the community behind them and attract the support of funding agencies. In order to maintain momentum, these initia-tives need to define achievable, tangible deliverables within a broader strategic re-search plan that is both scientifically excit-ing and societally relevant. Those deliver-ables need to leverage existing activities to the extent possible.

There was a clear consensus at the work-shop that there exists a notable gap be-tween scientific communities, as most peo-ple knew only a small minority of the other participants. As discussed above, it seems apparent that progress in polar predict-ability will require crossing disciplinary boundaries to understand the feedbacks be-tween the troposphere and the stratosphere, ocean, land, and sea ice. In the discussions, it became evident that the nature of these feedbacks appears to be somewhat differ-ent in the two hemispheres, because of the different geometries, leading to rather dif-ferent scientific questions.

In the Arctic, the ocean is contained within a basin with a couple of entry/exit points, and sea ice covers the polar region, allow-ing a strong ice-albedo feedback. While there are certainly important dynamical processes — e.g., the export of sea ice through the Fram Strait depends on the position and strength of the Beaufort Gyre

— climate scientists tend to treat the Arctic primarily from a thermodynamic perspec-tive, focusing on budgets of heat and (in the ocean) salinity. Probably the most burning societal question is the rate of warming in the Arctic, as this has numerous local con-sequences, including those that relate to an ice-free summer-time Arctic. Whilst it is plausible that the most extreme model predictions of summer-time sea-ice loss are in fact our best predictions, and that the observed rate of decrease in summer-time sea-ice extent is well understood, the confi-dence we have in those statements needs to be greatly strengthened.

In the Antarctic, the ocean is annular, sea ice is largely seasonal, and the centre of the polar region is covered by land ice and ice shelves. While there are certainly features of interest arising from the longitudinal asymmetry of the Antarctic continent, the dominant climate structures are the circum-polar jets in the atmosphere and the ocean, and climate scientists tend to treat the Ant-arctic primarily from a dynamical perspec-tive, focusing on eddy momentum fluxes and jet shifts. Furthermore, the largest ob-served changes in the Antarctic (which oc-cur in summer) are thought to be associated with the stratospheric ozone hole, reinforc-ing this dynamical perspective. On the other hand, the basic mechanisms for po-lar amplification (sea ice-albedo feedback, enhanced atmospheric latent heat flux) also exist in the Antarctic but are being delayed by deep ocean heat uptake, although it is unclear how well climate models represent this delay. Probably the most burning soci-etal question is what is the true response of the ocean circulation to the strengthening and poleward shift of the tropospheric jet, and how will this change in the future as the ozone hole recovers while GHG con-centrations continue to increase, as this has implications for Southern Ocean upwell-ing and carbon uptake, and possibly for the long-term stability of the West Antarctic ice shelf. In contrast to the situation in the Arctic, there is as yet no plausible explana-tion for the observed increase in Antarctic sea-ice extent, which remains a major sci-entific puzzle.

These are, of course, just the current ques-tions, but we can be sure that they will re-main “grand challenges” for some years yet, and furthermore that answering them (and contrasting the different behaviour of the two hemispheres) will advance our un-

(January 11, 2011 / 08:56:50)

69967-1newsletter36-v2_New_p18.pdf .1

19

derstanding of the fundamental processes and feedbacks underlying polar predict-ability. At the same time, a number of gen-eral issues and opportunities were identi-fied which apply to both poles:(i) A better understanding of seasonal

predictability, not only for its societal benefits but also for understanding the seasonality of longer-term variabil-ity and changes. The WCRP’s Working Group on Seasonal to Interannual Pre-diction (WGSIP) has the infrastructure to perform prediction studies but needs the expertise of polar scientists to inter-pret the results of those studies in polar regions and design new experiments.

(ii) A better understanding of decadal variability and its partitioning be-tween internally generated and exter-nally forced components. The WCRP’s Working Group on Coupled Modelling (WGCM) has defined a set of coordi-nated experiments focusing on the near term (i.e., several decade) time horizon within its CMIP5 activity, which will provide a large archive of model simu-lations that can be analysed from this perspective.

(iii) Improved initial state estimates. Po-

tential improvements in existing obser-vations (or their availability) need to be identified for action by the relevant agencies; coupled assimilation systems including snow and sea ice need to be developed, in collaboration with weath-er prediction centres who are wrestling with this issue as part of their efforts to improve polar weather prediction; and there needs to be a better understanding of the sensitivity of polar predictability on decadal time scales to initial-state er-ror in the ocean, to guide ocean observa-tional network design.

(iv) A better understanding of potential predictability. The value of a “perfect model” methodology hinges entirely on how realistic the model is. In cases where models have some basic cred-ibility, this approach can be exploited to determine where the predictability lies. In other cases, key model processes that are holding back progress need to be identified for a targeted effort at im-provement.

The conclusion of the workshop was that a cross-cutting WCRP initiative was needed in the area of polar predictability,

whose first action would be to hold a fo-cused meeting in about six months’ time, to develop a detailed implementation plan concerning the above issues. In developing such a plan it will of course be necessary to engage and partner with other relevant research bodies. It was felt that although there were important differences between the Arctic and Antarctic that could lead to differences in priorities, there were also considerable scientific and logistical ben-efits to be obtained by considering the two poles in parallel. Therefore it was suggest-ed that there should be a single initiative, but with distinct Arctic and Antarctic foci.

Acknowledgements

The workshop received substantial finan-cial support from the Norwegian Research Council and the WCRP, SPARC and CliC, and was also sponsored by the Bjerknes Centre, University of Bergen and the City of Bergen. We are extremely grateful to Beatriz Balino of the Bjerknes Centre for coordinating the excellent local arrange-ments, which made for a very productive meeting.

Report on the SPARC DynVar Workshop 2 on Modelling the Dynamics and Variability of the Stratosphere-Troposphere System

3-5 November 2010, Boulder, USA

E. Manzini, MPI-M, Hamburg, Germany (elisa.manzini@zmaw.de)N. Calvo, NCAR, USA and Universidad Complutense de Madrid, Spain (calvo@ucar.edu)E. Gerber, Courant Institute, USA (gerber@cims.nyu.edu)M. Giorgetta, MPI-M, Hamburg, Germany (marco.giorgetta@zmaw.de)J. Perlwitz, CIRES, University of Colorado and NOAA/ESRL PSD, Boulder, USA (judith.perlwitz@noaa.gov)L. Polvani, Columbia University, USA (LMP@columbia.edu)F. Sassi, Naval Research Laboratory, USA (fabrizio.sassi@nrl.navy.mil)A. Scaife, UK Met Office, UK (adam.scaife@metoffice.gov.uk)T. Shaw, Columbia University, USA (tas2163@columbia.edu)

The 2nd Workshop of the Stratospheric Pro-cess and their Role in Climate (SPARC) DynVar Activity took place in Boulder, Colorado, USA, 3-5 November 2010. The workshop was hosted by the National Oceanic and Atmospheric Administration (NOAA) Earth System Research Labora-tory’s (ESRL) Physical Sciences Division

in collaboration with the Cooperative Insti-tute for Research in Environmental Scienc-es (CIRES) at University of Colorado, and was held at NOAA ESRL David Skaggs Research Center. CIRES, NOAA, includ-ing NOAA’s Modeling, Analysis, Predic-tion and Projection (MAPP) Program, the European Commission COMBINE Inte-

grating Project, and the SPARC project of the World Climate Research Programme (WCRP) kindly provided support for the workshop. A special thanks to the many people at NOAA and CIRES involved in the organization of the excellent local ar-rangements for the Workshop at NOAA ESRL.

(January 11, 2011 / 08:56:52)

69967-1newsletter36-v2_New_p19.pdf .1

20

The SPARC DynVar Workshop 2 attracted 68 participants from 11 countries: USA (35), Canada (8), United Kingdom (7), Japan (6), Germany (4), France (3), Den-mark (1), Israel (1), Italy (1), Norway (1), Spain (1). The workshop consisted of 11 invited and 41 contributed presentations (11 orals and 30 posters) and was opened by a keynote presentation by Susan Solo-mon. Forty-five abstracts were submitted to the workshop, although submission of abstracts was not compulsory. The rela-tively large number of submitted abstracts indicates a growing interest in the role of stratospheric dynamics and variability on the climate system. Poster sessions were all well attended. Lunch and coffee breaks held on site were intensively used for in-formal discussions. A total of 5 hours was dedicated to discussing the core goals of the DynVar Activity, including difficulties and opportunities for those in the SPARC community, with most of the discussion fo-cused on the role of stratospheric dynami-cal processes in the Earth system.

The goals of the DynVar Activity are to determine the dependence of the mean climate, climate variability, and climate change on stratospheric dynamics as rep-resented in climate and Earth system mod-els. Since the first DynVar Workshop (held in Toronto, Canada, 27-28 March 2008), a number of new studies contributing to our knowledge on how stratospheric represen-tation affects climate simulated by models have appeared in the literature. In part, because of these advancements, a num-ber of climate modelling groups are now planning to undertake the Coupled Model Intercomparison Project Phase 5 (CMIP5) experiments with models that include a well-resolved stratosphere. The interest in models with a well-resolved stratosphere has also led to the Stratosphere resolving Historical Forecast Project (SHFP), part of the WCRP’s Working group on Seasonal to Interannual Prediction (WGSIP) cross cutting activity with the Climate Variabil-ity and Predictability project (CLIVAR), aimed at quantifying improvements in ac-tual predictability by initializing and re-solving the stratosphere in seasonal fore-cast systems.

The 2.5 day workshop provided a forum for: • Presenting new works on key areas cen-

tral to the Activity such as the influence of the stratosphere on the tropospheric

circulation, the ocean circulation via air-sea interactions, and on snow and sea-ice fields; the role of the strato-sphere in the tropospheric circulation response to climate change; and the mechanisms for two-way stratosphere-troposphere coupling;

• Assessing the status of the SHFP and CMIP5 runs with models with a well-resolved stratosphere.

• Discussing how to best analyse, make full use, and exchange knowledge from the data generated by the SHFP and CMIP5 runs, with the role of the strato-sphere as the focus.

The workshop agenda was organised based based on time scales: Presentations on in-terannual and shorter time scales, includ-ing discussion on the SHFP, occupied the first day, while the second and third days were dedicated to decadal and centennial time scales, and CMIP5 models and experi-ments.

The first day of the workshop started with a welcome by J. Perlwitz and an introduc-tion of the DynVar activity and workshop goals by E. Manzini. In her opening key-note presentation, S. Solomon reviewed a number of challenges that the climate community is facing, such as understand-ing the reasons for decadal variations in stratospheric water vapour, modelling the chain of processes in the tropical atmo-sphere that may bring meteorological sig-nals originating in the lower atmosphere to the stratosphere, the importance of the location of the lid of a model, and the ac-curate representation of stratospheric pro-cesses in models. She acknowledged the role of variability, reviewed the role of the stratosphere in connecting changes occur-ring in the Antarctic region to global cli-mate change, and presented new results on temperature trends in the UTLS. J. Perl-witz reported on the WRCP Workshop on Seasonal to Multi-decadal Predictability of Polar Climate held in Bergen, the week prior (25-29 October 2010). Topics of rel-evance to DynVar were the sources of po-tential predictability reviewed during the workshop, especially those associated with stratospheric processes, and the establish-ment of both an Arctic and an Antarctic Initiative. She also presented the SHFP-WGSIP activity on behalf of A. Sciafe, and called for leadership from the DynVar group in the analysis of the SHFP runs. The SHFP runs are seasonal hindcast experi-

ments, generally carried out with coupled atmosphere-ocean-sea-ice models, which are also high-top models. J. Scinocca pre-sented the CCCma contribution to SHFP, although in this case, the high-top seasonal hindcasts were performed with imposed sea surface temperatures (SSTs) and sea-ice concentrations (SICs).

M. Baldwin reviewed methodologies to diagnose stratosphere-troposphere cou-pling in both observations and simulations. A key issue is how to define a climatology in a changing climate. By defining a slowly varying climatology with specific statisti-cal properties, the resulting Annular Mode (AM) indices have no trend by definition – meaning that the climatology will change but the annular mode of variability will not. Baldwin suggested using daily zonal-mean geopotential to define the AM from climate model outputs, after removing its daily global-mean, a slowly varying trend and the seasonal cycle to define the anomalies. D. Waugh reported that CCMVal-1 and CCMVal-2 have demonstrated the advan-tage of the multi-model evaluation strat-egy, combined with model grading, over a range of diagnostics for the identification of deficiencies and systematic biases in chemistry-climate models. These activities have also led to quantifiable improvements in some particular models in the subjects of transport, Cly abundance and tropical tropopause temperatures. However, the methodology of model grading has its own limitations, such as the robustness of the metrics and the determination of the uncer-tainties in the observations used for com-parison. Of particular relevance to DynVar are the results of Chapter 10 of the SPARC CCMVal-2 report (Baldwin et al., 2010), which demonstrates that the CCMVal-2 models, which generally have a better-resolved stratosphere, perform better than AMIP CMIP-3 models in the stratosphere and perform equally well, if not better, in the troposphere. The reported CCMVal di-agnostic tool appeared to be of interest to many analysts and model developers.

The contributed talks of the first day in-cluded oral and poster presentations on a variety of topics, including the role of the stratospheric ozone on the medium-range weather forecast (M. Deushi), the role of linear interference in the annular mode re-sponse to tropical forcing (P. Kushner), wave forcing of the QBO (J. Anstey), the evaluation of the stratosphere in seasonal

(January 11, 2011 / 08:56:54)

69967-1newsletter36-v2_New_p20.pdf .1

21

forecast models (A. Butler and A. May-cock) and on the factors controlling decor-relation time scales in the lower strato-sphere (P. Hitchcock).

N. Butchart opened the second day with a talk on climate change and stratosphere-troposphere interactions, and pointed out that the effect of stratospheric changes on surface may not be limited to the impact of Antarctic ozone depletion and recov-ery. According to the multi-model study reported, the inter-comparison of the at-mospheric response to 4xCO2 in low- and high-top models showed that stratospheric climate changes may contribute substan-tially to changes in storm tracks, sea level pressure and precipitation in the Northern Hemisphere during winter. The fact that the impact of a well-resolved stratosphere stands out in the reported multi-model comparison suggests that results are robust, despite widely differing parameter settings and schemes in the high- and low-top mod-els. However, a limitation of the reported work is the specification of the SSTs and SICs, disabling any air-sea interactions in the high top models, such that the climate is slaved to the imposed SSTs and SICs. It is therefore paramount to call for a simi-lar analysis, with high- and low-top atmo-sphere-ocean general circulation models (AOGCMs).

Discussion of the status of the develop-ment of AOGCMs with a well-resolved stratosphere followed. In most cases, these models are high-top versions of low-top models. C. Cagnazzo reported results from the CMCC, IPSL-CM5, and MPI models. These three modelling systems, together with EC-Earth presented by S. Yang, and the METO&UK Universities presented by S. Hardiman, participate in the COMBINE European Integrating Project that aims to develop the next generation of Earth Sys-tem Models by including components such as a dynamical stratosphere. The model de-scriptions and status of the CMIP5 simula-tions were given for the GFDL CM3 model (J. Austin), MIROC-ESM (S. Watanabe), WACCM (D. Marsh), GEOS-5 (S. Paw-son), and MRI (K. Shibata). There were therefore 10 high-top model systems pres-ent at the DynVar workshop, with at least three models (EC-Earth, METO&UK Universities and WACCM) that have low-top counterparts. At the time of the work-shop, pre-industrial control simulations were completed (or close to completion)

for all the models. Some centres (e.g., GFDL) were finishing the majority of the core CMIP5 long-term experiments (pre-industrial control run, 1850-2005 historical run, RCP4.5 and RCP8.5 runs). GEOS-5 and the MPI model were the only high top AOGCMs planning to run the CMIP5 decadal prediction experiments. At least three model systems (GFDL CM3, MPI and MIROC-ESM) will also run with CO2 emissions, requiring modules for the land and ocean carbon cycle. Interactive at-mospheric chemistry was included in at least three model systems (GFDL CM3, MIROC–ESM and WACCM). Differ-ent modelling groups were using similar types of diagnostics to analyse some of the most current topics of research in the troposphere-stratosphere region. These topics include ENSO signals in the trop-ics, ENSO teleconnections in the Northern Hemisphere, the Atlantic meridional over-turning circulation (AMOC), simulation of the QBO and its forcing, and changes in tropical upwelling due to climate change, and decadal variability in water vapour. This potential for collaborative studies is precisely what initiatives such as DynVar are meant to address.

The last 2 sessions of contributed oral and posters presentations featured, the relative role of ozone and dynamical trends (and their model biases) in the southern polar stratospheric temperature trends (N. Cal-vo), the relationship between stratospheric ozone and Antarctic sea-ice trends (M. Sigmond), changes in the reflective down-ward coupling associated with ozone de-pletion (N. Harnik), evidence of coupling between the North Annular Mode and low frequency AMOC variability, suggesting a connection between stratospheric variabil-ity and variation in deep ocean temperature variations (J. Kim), and the dynamical en-hancement of the equator to pole contrast in tropopause height, by more than a fac-tor 2 compared to the radiative equilibrium solution (T. Birner). Posters also covered a wide range of topics that were both di-rectly relevant to DynVar, or indicate fruit-ful interactions between DynVar and other SPARC activities. For example, the con-nection with gravity waves was highlighted as a prerequisite to calculate accurate mo-mentum budgets in the stratosphere, which is a focus of the SPARC Gravity Wave Activity and is also relevant to some of the DynVar topics. Different studies on chang-es in atmospheric composition, in particu-

lar water vapour and ozone also indicated that DynVar could exploit interactions with CCMVal. Similarly, the role of dynamics in mediating the solar cycle signal from the stratosphere to the surface indicates an interaction with SOLARIS. Other posters were more specific to the DynVar objec-tives, covering for example the role of re-solved planetary waves generated in asso-ciation with tropical warm pools of SSTs.

Presentation sessions were complemented by discussion sessions dedicated to ad-dressing how the SPARC community could make use of the opportunities generated by international activities such SHFP and CMIP5. The final session on Friday was dedicated to consolidating future efforts and plans. A number of activities were pro-posed and are summarised here: 1. Evaluate the feasibility of writing “news

and views” papers on the role of strato-spheric dynamics on tropospheric cli-mate (Edwin Gerber, Natalia Calvo and Tiffany Shaw)

2. Evaluate the feasibility of writing a re-view paper on the changes occurring in the Antarctica region, focusing on the effects of ozone depletion on the climate system, including the ocean carbon flux-es (Judith Perlwitz)

3. Coordinate two synthesis papers on the CMIP5 runs: (i) Multi-model high-top model comparison of stratospheric cli-mate, variability and change (Andrew Charlton-Perez) and (ii) Multi-model high-top / low-top comparison focused on surface climate, variability and change (Elisa Manzini)

4. Establish research groups to foster anal-ysis of the SHFP and CMIP5 archives, towards a workshop in mid-2012, to be proposed to SPARC at the next Scien-tific Steering Group (SSG) meeting.

While each modelling centre has planned its own papers on the validation and/or novel applications of the new high-top AOGCMs, it is envisaged that a number of studies may explore the simulations that will become available through the SHFP and CMIP5 archives. DynVar is seen in this respect as a facilitator, fostering collab-orative analysis on the role of stratospheric dynamics in the climate system, and on the implications of stratosphere-troposphere dynamical coupling for the prediction of variability and change of the climate sys-tem at all time scales. This stage of Dyn-Var is foreseen to last at least for the next

(January 11, 2011 / 08:56:56)

69967-1newsletter36-v2_New_p21.pdf .1

22

The SPARC Data Initiative is the newest SPARC activity and was launched during the 2009 SPARC SSG meeting in Kyoto, Japan. Its aim is to compare data sets of vertically resolved chemical trace gas ob-servations obtained from different satel-lite instruments, and to provide a “user’s guide” for the use of such data in different applications, such as model-measurement comparisons or empirical studies of strato-spheric climate and variability.

About 10 years ago, the GCM-Reality Inter-comparison Project for SPARC (GRIPS) found that there was considerable uncertainty in its model inter-comparison of dynamics and radiation arising from the fact that different observed data sets often delivered conflicting results. Accordingly, a middle atmosphere climatology study was initiated by SPARC, which compared the available meteorological data prod-ucts in terms of various aspects including mean biases, seasonal cycle, variability, and long-term changes. No data set was problem-free, and all data sets were found to have both strengths and weaknesses. The findings were published in the SPARC Re-port No. 3 (2002), which provided some-thing of a user‘s guide to the data.

The same sort of situation was faced in

the SPARC CCMVal (Chemistry-Climate Model Validation) project (Eyring et al., 2005) for chemical trace gas measure-ments. While ozone and water vapour measurements are the subject of specific SPARC activities, there is no equivalent activity for other chemical trace gases. Yet these gases play an essential role in the ozone budget, and, together with age of air (a derived product), provide tracer in-formation on atmospheric transport; a topic extensively analysed in the recent CCMVal Report. There are a variety of trace gas data sets available and a user cannot easily de-termine which is the most reliable for any particular application. While comparison of different measurements is often done as part of instrument validation studies, this information is not readily available to us-ers. Moreover, the data sets are not always available in a standard data format, or with appropriate documentation. The result was that for the CCMVal inter-comparison, dif-ferent observational data sets were used by different people, and scores based on model metrics were highly dependent on the data set employed. The SPARC CCMVal report therefore identified the need for an assess-ment of the available data sets of chemical trace gases analogous to what was done in SPARC Report No. 3 for the meteorologi-

The SPARC Data Initiative

M. I. Hegglin, University of Toronto, Canada (michaela@atmosp.physics.utoronto.ca)S. Tegtmeier, IFM Geomar, Germany (stegtmeier@ifm-geomar.de)

cal data sets. A specific recommendation states ‘A systematic comparison of exist-ing observations is required in order to un-derpin future model evaluation efforts, by providing a more accurate assessment of measurement uncertainties’.

Responding to this recommendation, the SPARC Data Initiative aims at assessing and consolidating our knowledge of cur-rent and past space-based observations of chemical trace gas species in the upper tro-posphere, the stratosphere, and the lower mesosphere. Both long-lived (O3, H2O, N2O, CH4, CFCs, SF6, HF, NOy, Bry, HCl and CO) and short-lived trace gas species (NO, NO2, NOx, HNO3, HNO4, N2O5, ClO-NO2, BrONO2, ClO, HOCl, BrO, OH, HO2, CH2O), as well as aerosols will be assessed. The goal of the project is to assemble and compare climatologies derived from obser-vations, to identify differences between the data sets, and to provide expert judgment on the source of those differences. The re-sults will be documented in a new SPARC report, which will also provide essential knowledge of the measurement and retriev-al techniques used. The report will compare quantities including zonal mean climatolo-gies, seasonal evolution, and interannual variability of the chemical species.

two years. To this end, DynVar “Research Groups” are being established on a number of topics raised at the workshop. Proposed research groups proposed include: Ant-arctica: From Ozone to Carbon; Surface climate, variability and change; Sudden Stratospheric Warming; ENSO and QBO; AMOC and PDO; Water vapour; Annular Modes / Stratospheric memory; QBO and tropical waves; Tropopause and External forcing (volcanic). Concerning the solar external forcing and gravity waves top-ics, we note that the SPARC SOLARIS and Gravity Wave activities already exist to study these areas. Collaboration with CCMVal on a variety of issues is also en-visioned. To foster the collaboration with CLIVAR, Amy Butler and Adam Scaife have volunteered to be the DynVar contacts on the SHPF project. Research Groups and

their contacts will be posted on the DynVar web site (http://www.sparcdynvar.org/).

To recognize the engagement of a number of new people at the core of the SPARC DynVar Activity, the DynVar Committee has been restructured: Amy Butler, Nata-lia Calvo, Andrew Charlton-Perez, Edwin Gerber, Tiffany Shaw and Shingo Watan-abe are welcomed as new members; whileJudith Perlwitz, Lorenzo Polvani and Fab-rizio Sassi will remain involved as ex-offi-cio members.

We would like to note in closing that the larger than expected participation in the workshop clearly highlights the need for a forum of discussion on stratospheric dy-namics in the interim period between the last SPARC General Assembly in 2008 and

the next one in 2014, possibly reaching outto CLIVAR and CliC (The Climate and Cryosphere project). Given the above and the proposed 2-year time scale for the fos-tering the analysis of the SHFP and CMIP5 runs, we propose to hold a one week Work-shop in spring/summer 2012 with SPARC DynVar and CLIVAR/SHFP, and possibly CliC.

References

Baldwin, M. et al., 2010: Effects of the stratosphere on the troposphere, in the SPARC Report on the Evaluation of Chem-istry-Climate Models, V. Eyring, T.G. Shepherd, D.W. Waugh (Eds.), SPARC Report No. 5, WCRP-132, WMO/TD-No. 1526.

(January 11, 2011 / 08:56:58)

69967-1newsletter36-v2_New_p22.pdf .1

23

At this point, a core team of instrument sci-entists and data analysts has been defined, and has started assembling trace climatolo-gies from different satellite instruments and evaluating the available data products. The core team consists is listed in Table 1. The instruments represented by the core team members and by confirmed contributors to the SPARC Data Initiative currently in-clude ACE-FTS, ACE-MAESTRO, Aura-MLS, GOMOS, HALOE, HIRDLS, LIMS, MIPAS, Odin OSIRIS, Odin SMR, SAGE, SCIAMACHY, SMILES, TES, and UARS-MLS. The core team met for a first success-ful workshop at the International Space Sci-ence Institute (ISSI) in Bern, Switzerland from 22-26 November 2010, thanks to gen-erous support for local costs provided from ISSI through their international team pro-gramme (see our ISSI team website http://www.issibern.ch/teams/atmosgas/index.html/Welcome.html) and for travel from the WCRP. A workshop report will appear in the July issue of the SPARC newsletter. There will be a second and possibly even a third workshop, leading to the completion of a SPARC report.

and retrieval techniques, and validation activities. In particular, the initiative will help to identify priorities for reprocessing existing data or enhanced validation efforts with the active support of the measurement groups. In addition, the project will identify measurement gaps, which could motivate and provide community support for future missions.

The SPARC Data Initiative

M. I. Hegglin, University of Toronto, Canada (michaela@atmosp.physics.utoronto.ca)S. Tegtmeier, IFM Geomar, Germany (stegtmeier@ifm-geomar.de)

The timeliness of the SPARC Data Initia-tive is supported by the fact that the last few decades represent something of a golden age of stratospheric composition measurements, and it is unlikely that the stratosphere will be as well observed in the future. This initiative will help to capture and summarize existing knowledge on cur-rent and recent instruments, measurement

Table 1: List of core members.John Anderson Hampton University, USA HALOE, UARSSamuel Brohede Chalmers University, Sweden OSIRIS, OdinLucien Froidevaux Jet Propulsion Laboratory, USA MLS, Aura-MLS, UARSBernd Funke Instituto de Astrofísica de

Andalucía, SpainMIPAS, ENVISAT

Michaela Hegglin (co-lead) University of Toronto, Canada data-analysisAshley Jones University of Toronto, Canada ACE-FTS, SCISAT-1Erkki Kyrölä Finish Institute for Meteorology,

Finland GOMOS, ENVISAT

Jessica Neu University of California Irvine, USA TES, AuraAlexei Rozanov University of Bremen, Germany SCIAMACHY, ENVISATSusann Tegtmeier (co-lead) IMF-GEOMAR, Germany data-analysisMatthew Toohey IMF-GEOMAR, Germany data-analysisJoachim Urban Chalmers University, Sweden SMR, Odin

Thomas von Clarmann Karlsruhe Institute of Technology, Germany

MIPAS, ENVISAT

Kaley Walker University of Toronto, Canada ACE-FTS & MAESTRO, SCISAT-1

Program of the Antarctic Syowa MST/IS Radar (PANSY) K. Sato, The University of Tokyo, Japan (kaoru@eps.s.u-tokyo.ac.jp)M. Tsutsumi, National Institute of Polar Research, Japan (tutumi@nipr.ac.jp)T. Sato, Kyoto University, Japan (tsato@kuee.kyoto-u.ac.jp)T. Nakamura, National Institute of Polar Research, Japan (nakamura.takuji@nipr.ac.jp)A. Saito, Kyoto University, Japan (saitoua@kugi.kyoto-u.ac.jp)Y. Tomikawa, National Institute of Polar Research, Japan (tomikawa@nipr.ac.jp)K. Nishimura, National Institute of Polar Research, Japan (knish@nipr.ac.jp)H. Yamagishi, National Institute of Polar Research, Japan (yamagisi@nipr.ac.jp)T. Yamanouchi, National Institute of Polar Research, Japan (yamanou@nipr.ac.jp)

Introduction The polar regions play an important role in the Earth’s climate system. They are the exit regions of the material circulation in the stratosphere, and the entry (exit) regions during the summer-time (winter-time) in the mesosphere. This material circulation is essentially wave-driven and maintains a thermal structure of the middle atmosphere far from that expected by radiative balance. The resulting low temperatures in the summer upper mesosphere and winter lower stratosphere lead to conditions under

which polar mesospheric clouds (PMC) and polar stratospheric clouds (PSC) can form, respectively. The PSCs serve as an environment for producing the conditions that can lead to catalytic destruction of ozone during Antarctic spring, forming the ozone hole. PMCs are a phenomenon that was first reported late in the 19th century and were considered to have appeared due to the changing climate, after the Industrial Revolution. The PMCs have also recently been observed even at mid-latitudes, such as those observed over Paris in the summer of 2009, which are still fresh in

our memory. Thus, PMCs can be directly related to human activity and considered “the canaries in a coal mine” of the Earth climate system. Therefore, the monitoring and study of these phenomena is an important tool in detecting climate change, and in understanding the Earth system.

Another interesting polar phenomenon is the high albedo and high elevation of the Antarctic continent, which can cause strong downslope (katabatic) winds in the coastal region. These winds are an important source of gravity waves in the

(January 11, 2011 / 08:56:59)

69967-1newsletter36-v2_New_p23.pdf .1

24

Antarctic. Moreover, the polar atmosphere is different from lower latitude regions because of possible strong energy inputs from geospace due to the orientation of the Earth’s magnetic field lines, which are directly connected to the plasma sheet there.

Despite the importance of the southern polar atmosphere, observational studies have thus far been restricted due to the harsh physical conditions on the Antarctic continent. In order to address this deficiency, to explore the physics of these unique phenomena, and to study the quantitative effects of the polar atmosphere on the Earth’s climate, a Mesosphere–Stratosphere-Troposphere/Incoherent Scatter (MST/IS) radar (a VHF clear-air Doppler radar) will be installed at Syowa Station (69°S, 40°E) in early 2011; the first of its kind in the Antarctic (see Figure 1; colour plate III). We call this radar the PANSY radar after the name of the project. “PANSY” is a flower name coming from a French word “penser” which means “to think”.

Radar system

The PANSY radar is a monostatic pulse Doppler radar operating at 47 MHz. Its antenna is a circular active phased array 160 m in diameter, consisting of 1045 three-element crossed Yagi antennas. The half-power beam width is 2.4 degrees, and the beam direction can be pointed to any specified direction within 30 degrees from the zenith. Each of the 1045 antennas is equipped with a solid-state transmit-receive module with 500 W peak output power, and the total output power is 520 kW. Fundamental parameters of the radar are given in Table 1.

The basic concept of the PANSY radar is to have a sensitivity comparable to that of the MU radar in Shigaraki, Shiga, Japan, which covers the height region of 1.5-600 km (Fukao et al., 1985). One of the major technical limitations in designing the

PANSY radar is its power c o n s u m p t i o n . While the MU radar with its class-AB power amplifier consumes 230 kW, the maximum available power for the PANSY radar at Syowa

Station is less than 100 kW. Since the sensitivity of an atmospheric radar is determined by the product of the total output power and the antenna’s effective area, we reduced the output power to half of that of the MU radar, and doubled the antenna area. In order to further reduce the power consumption, we developed a new class-E amplifier, with a total power efficiency exceeding 50%. As a consequence, we achieved a total power consumption of 75 kW, including the power needed for the digital signal processing system. This power-efficient design enabled us to avoid the use of a cooling fan, resulting in a robust transmit-receive module free from any moving components, ideal for the long-term operation in the severe weather conditions in Antarctica.

Another important design issue is the construction of the radar at Syowa Station in the Antarctic. The construction must be performed during a short summer period of about 2 months, and by untrained members of the Antarctic research expedition, most of whom are scientists (not technicians). Also, due to restrictions by the Antarctic Treaty, the rough ground cannot be altered, preventing the use of heavy construction vehicles. We have thus made our best effort to minimise the weight of the antenna and the transmit-receive module. In addition, the three-element crossed Yagi antenna has to withstand maximum continuous wind speeds of up to 65 m/s. After examining several test antennas at Syowa Station, we developed the current design, with a weight of 12 kg. The transmit-receive module is also designed to have minimum weight (18 kg) and cross-section. The supporting mast is inserted into a mounting hole of 130 mm diameter and 1 m depth, and the antenna is attached to its top together with the transmit-receive module. The height of each antenna is variable, in a range of about 2 m, depending on the ground surface condition, and the phase of the

radiated signal is electronically adjusted to compensate for the height difference. The antennas are arranged to form a grid of equilateral triangles with 4.5 m intervals, which corresponds to a wavelength of 0.7 m. Nineteen antennas constitute one group in a hexagonal shape, and a divider/combiner module is located at near the centre of each group. The entire antenna array consists of 55 hexagonal groups. The cables for the RF signal, the control signal, and the DC power supply spread from the operating building located next to the array field.

The transmitted pulse length is variable with minimum length of 0.5 microseconds, corresponding to a height resolution of 75 m. In order to increase sensitivity, Spano codes, which are an extension of complementary codes, are employed. Because of the non-linear nature of the class-E amplifier, each bit of the pulse code has to be separated in time. We thus use a train of pulses with an interval equal to the pulse length. The antenna beam direction can be switched by electronically controlling the phase of each antenna module for each transmitted pulse. The basic observation scheme for the troposphere and stratosphere will be 5 beam directions (vertical, north, east, south, and west) with a one-minute time resolution. The received signal is processed by an array of 55 digital receivers. Each group of the antenna is connected separately to a digital receiver, so that imaging observations of atmospheric turbulence, as well as the ionospheric irregularities can be performed.

One special nature of the Antarctic VHF radar is the existence of a strong coherent echo due to ionospheric field-aligned ir-regularities (FAI) of auroral origin. These echoes are observed in a direction perpen-dicular to the Earth’s magnetic field, which has a declination angle of about 70 degrees near Syowa Station. As the main antenna array cannot be pointed to an elevation an-gle of 20 degrees, a one-dimensional array of 24 Yagi’s is arranged around the outer edge of the main circular array. The output of the array can be connected to each of 8 digital receivers devoted to the FAI ob-servations. The same peripheral array will be used to suppress possible interferences of FAI echoes with weak incoherent scat-ter echoes observed by the main array. A special adaptive antenna algorithm will be developed to cope with this problem.

The first stage of the radar construction will

Table 1: Basic parameters of the PANSY radar.System Pulse Doppler radar

Active phased array systemCentre Frequency 47 MHzAntenna A quasi-circular array consisting of 1045 crossed

Yagi antennas. Diameter about 160 m (18000 m2)Transmitter 1045 solid-state TR modules

Peak Power: 520 kW Receiver (55+8) channel digital receiving systems

Ability of imaging and interferometry observationsPeripheral 24 antennas for E-layer FAI observation

(January 12, 2011 / 12:00:11)

69967-1newsletter36-v2_New_p24rev.pdf .1

25

take place at Syowa Station in December 2010, and is scheduled for completion in March 2011. Tropospheric observation will be started with a small system of 57 antennas at this stage. The full system will start operation in March 2012. For the purposes of training the radar operators and developing the observation software, including the adaptive clutter rejection algorithm, a small system consisting of 22 antennas was constructed in the MU Observatory, as shown in Figure 2.

Backscattering of the atmospheric radar is caused by fluctuations in the refractive index of the air. More precisely, the backscattering is the Bragg scattering from the wave number component of the fluctuations that are half the radar wavelength (3.2 m for the PANSY radar). The main contribution to fluctuations in the index of refraction in the lower and middle atmosphere is disturbances in the background gradient of air density with height caused by atmospheric turbulence. Fluctuations in water vapour also play an important role in the lower troposphere. In addition to the background wind, the intensity of the atmospheric turbulence can be measured in this height region. As the air density decays exponentially with height, the fluctuations, and thus the echo power, also decay. The mean lapse rate of the echo power in the mid-latitude lower stratosphere is about 2-3 dB/km. The maximum observation height of the PANSY radar in the stratosphere is expected to be about 25 km.

Ionization becomes the major contribution to the refractive index above about 50 km. The advantage of using the lower VHF band is that the turbulent eddies still have substantial magnitude at the size of half the radar wavelength at mesospheric heights. Coherent scattering echoes due to turbulence and enhanced by the ionization are observed in the daytime mesosphere (60-80 km). During the summer season, enhanced coherent echoes, called Polar Mesosphere Summer Echo (PMSE), are also observed.

Above about 100 km up to about 500 km, incoherent scattering by ionospheric electrons can be observed by the PANSY radar. Electron and ion density, their temperatures, and the ion drift velocity are derived by analysing the echo power spectra. At around 80-105 km, spontaneous echoes from meteor trails can also be used

to measure the background wind.

Scientific targets

The basic observation mode of the PANSY radar for the neutral atmosphere provides vertical profiles of three-dimensional wind vectors over a wide height range in the troposphere, stratosphere and mesosphere. Vertical resolution is at best 75 m, but is usually 150 m for the troposphere and stratosphere and 300 m for the mesosphere. Horizontal resolutions corresponding to the half-power beam width 2.4 degree are 840 m at the height of 20 km and 3.14 km at 75 km. Various atmospheric phenomena and dynamical processes are examined by continuous observation with the fine vertical and horizontal resolutions (e.g., polar lows causing severe snow storms, tropospheric circulation associated with katabatic winds, fine structure of the tropopause, stratopause and mesopause, sudden stratospheric warming, polar vortex break-up, medium-scale Rossby waves trapped at the edge of the polar vortex, dynamics of PSC and PMC, atmospheric turbulence, and atmospheric gravity waves). The response of the neutral atmosphere to the injection of high energy particles from the magnetosphere is also an important topic of the PANSY project. Among these numerous themes, here we focus on possible research regarding gravity wave dynamics and PMC physics, which we consider the most important topics.

The spectra of atmospheric gravity waves are distributed over a wide range of frequency, and horizontal and vertical wavenumbers. Consequently, it is usually impossible to observe the whole gravity wave continuum by a single instrument

because of the observational filter problem (Alexander et al., 2010). Continuous observation with fine temporal resolution at one location by the MST radar allows us to extract gravity wave components by their high frequency nature using a time filter. The horizontal and vertical resolution of the MST radar is sufficiently high to detect gravity waves with even small horizontal and vertical scales. Thus, it can observe gravity waves in almost all horizontal and vertical wavenumber ranges if they have high ground-based frequencies, which covers a part of the spectral range that is invisible to conventional radiosondes and recent high-resolution satellite observations. Moreover, the ground-based frequency is conserved during wave propagation if the background field is steady. The conservation of frequency is a reasonable assumption compared with that of the vertical wavenumber since the mean wind usually has vertical shear. Thus, the wave force can be estimated by using the vertical profile of the momentum fluxes associated with gravity waves observed by the MST radar. Moreover, direct and accurate estimates of the momentum flux vector associated with atmospheric gravity waves are possible using the dual beam method (Vincent and Reid, 1983).

According to the gravity wave study by Yoshiki and Sato (2000) using operational radiosonde data, the characteristics of the atmospheric gravity waves at Syowa Station are similar to those at the other Antarctic Stations. Thus, it is expected that the nature of the waves observed by the PANSY radar will be representative of those over the Antarctic, despite the fact that observations are only at a single location. Recent work using high-resolution satellite observations (Ern et al., 2004; Alexander

Figure 2: The training system of the PANSY radar located at the MU Observatory, Shiga-raki, Shiga, Japan.

(January 11, 2011 / 08:57:03)

69967-1newsletter36-v2_New_p25.pdf .1

26

et al., 2008) and a gravity-wave resolving general circulation model (Watanabe et al., 2008; Sato et al., 2009) show that gravity waves have large energy in the high latitude regions of the Southern Hemisphere in austral winter. Generation and propagation of such gravity waves, as well as their momentum deposition to higher latitudes, will be elucidated by the PANSY radar in combination with satellite and model data. Collaboration with the radars at the other stations such as the ST radar at Davis station (Australia) is also important to examine the locality of wave characteristics.

Another interesting topic is clouds in the polar middle atmosphere, in particular, PMCs. It is well known that a strong coherent echo (PMSE) is observed from the polar summer mesosphere. This echo is considered to be related to the PMCs (Cho and Roettger, 1997; Rapp and Luebken, 2004). The PMSE is observed by various kinds of radars with frequencies ranging from a few MHz to several hundred MHz. In addition to the PMSE, the PANSY radar can detect echoes from turbulence in the mesosphere. Thus, three-dimensional winds can be estimated regardless of the presence of PMCs. A Rayleigh lidar to be co-located at Syowa Station in early 2011, together with the PANSY radar, will be useful in clarifying the structure and evolution of PMCs/PMSE over the Antarctic. By imaging observations, the three-dimensional fine structure of PMSE and turbulence will be elucidated. A similar observational project is planned by MAARSY (Middle Atmosphere Alomar Radar System), installed at Andøya,

Norway (Latteck et al., 2010). The comparison of PMC characteristics in the Antarctic and Arctic atmosphere should be interesting. First, it is known that the PMSE is much weaker in the Antarctic than in the Arctic. In addition, recent studies (Becker and Fritts, 2006; Karlsson et al., 2009) have shown a possible link between the two hemispheres.

Global models for weather prediction and climate projection with relatively low resolution still have a cold bias in the winter polar stratosphere, probably in part because of unrealistic gravity wave parameterizations. This bias significantly degrades the predictions of the Antarctic ozone hole and its recovery because stratospheric temperature affects PSC volume. Quantitative understanding of polar atmospheric dynamics by adding the PANSY radar to the current observational network in combination with higher-resolution GCMs will reduce such model biases and contribute to the improvement of climate prediction. See the PANSY programme website for details: http://pansy.eps.s.u-tokyo.ac.jp

References

Alexander, M. J. et al., 2008: Global estimates of gravity wave momentum flux from High Resolution Dynamics Limb Sounder observations, J. Geophys. Res., 113, doi:10.1029/2007JD008807.

Alexander, M. J. et al., 2010: Recent developments in gravity-wave effects in climate models and the global distribution of gravity-wave momentum flux from ob-servations and models, Q. J. R. Meteorol. Soc., 136, 1103–1124.

Becker, E. and D. C. Fritts, 2006: Enhanced gravity-wave activity and interhemispheric coupling during the MaCWAVE/MIDAS northern summer program 2002, Annales Geophysicae, 24, 1175–1188.

Cho, J. Y. N. and J. Roettger, 1997: An updated re-view of polar mesosphere summer echoes: Obser-vation, theory, and their relationship to noctilucent clouds and subvisible aerosols, J. Geophys. Res., 102, 2001–2020.

Ern, M. et al., 2005: Absolute values of gravity wave momentum flux derived from satellite data, J. Geo-phys. Res., 109, 10.1029/2004JD0004752.

Fukao, S. et al., 1985: The MU radar with an active phased array system, 1. Antenna and power amplifi-ers, Radio Sci., 20, 1155-1168.

Karlsson, B. et al., 2009: Inter-hemispheric meso-spheric coupling in a comprehensive middle atmo-sphere model, J. Atmos. Solar-Terr., Phys., 41, 518-530.

Latteck, R. et al., 2010: MAARSY – the new MST radar on Andøya/Norway, Adv. Radio Sci., 8, 219-224.

Rapp, M. and F.-J. Luebken, 2004: Polar mesosphere summer echoes (PMSE): review of observations and current understanding, Atmos. Chem. Phys., 4, 2601–2633.

Sato, K. et al., 2009: On the origins of gravity waves in the mesosphere, Geophys. Res. Lett. 36, doi: 10.1029/2009GL039908

Vincent R. A. and I. M. Reid, 1983: HF Doppler mea-surements of mesospheric gravity wave momentum fluxes, J. Atmos. Sci., 40, 1321–1333.

Watanabe, S. et al., 2008: General aspects of a T213L256 middle atmosphere general circulation model, J. Geo-phys. Res. 113, doi:10.1029/2008JD010026.

Yoshiki, M. and K. Sato, 2000: A statistical study of gravity waves in the polar regions based on opera-tional radiosonde data, J. Geophys. Res. 105, 17995-18011.

Concordiasi is an international effort lead by Météo-France, and involving several research centres and polar institutes (Ra-bier et al., 2010). The project is organised around several observation campaigns in Antarctica, and its main scientific objec-tives are:• to improve the assimilation in numeri-

cal weather prediction (NWP) models of infrared radiances provided by IASI-like hyper-spectral space-borne sound-ers over icy surfaces;

• to enhance the representation of polar processes in numerical models, and in particular to improve the simulation of precipitation and clouds, so as to better describe the mass budget of ice sheets, as well as to provide observational con-straints on gravity-wave drag param-eterizations used in global circulation models;

• to advance our knowledge of micro-physical and dynamical processes in-volved in stratospheric ozone loss, and

Concordiasi: A Project Dedicated to the Polar Atmosphere

F. Rabier, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (florence.rabier@meteo.fr)A. Hertzog, Ecole Polytechnique, CNRS, France (Albert.Hertzog@lmd.polytechnique.fr)Ph. Cocquerez, Centre National d’Etudes Spatiales, France (philippe.cocquerez@cnes.fr)S. A. Cohn, National Center for Atmospheric Research, USA (cohn@ucar.edu)L. Avalonne, University of Colorado, USA (linnea.avallone@lasp.colorado.edu)T. Deshler, University of Wyoming, USA (deshler@uwyo.edu)J. Haase, Purdue University, USA (jhaase@purdue.edu)B. Brioit, Ecole Polytechnique, CNRS, France (brioit @ lmd.polytechnique.fr)F. Danis, Ecole Polytechnique, CNRS, France (francois.danis@lmd.polytechnique.fr)F. Vial, Ecole Polytechnique, CNRS, France (Francois.Vial@lmd.polytechnique.fr)A. Doerenbecher, Centre National de Recherches Météorologiques, CNRS and Mé-téo-France, France (alex.doerenbecher@meteo.fr)V. Guidard, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (vincent.guidard@meteo.fr)D. Puech, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (dominique.puech@meteo.fr)H. Cole, National Center for Atmospheric Research, USA (cole@ucar.edu)J. Fox, National Center for Atmospheric Research, USA (jfox@ucar.edu)T. Hock, National Center for Atmospheric Research, USA (hock@ucar.edu)D. Parsons, National Center for Atmospheric Research, USA (parsons@ucar.edu)J. VanAndel, National Center for Atmospheric Research, USA (vanandel@ucar.edu)L. Kalnajs, University of Colorado, USA (kalnajs@colorado.edu)C. Genthon, Laboratoire de Glaciologie et Géologie de l’Environnement, France (christophe.genthon@lgge.obs.ujf-grenoble.fr)

The 2011 NDACC Symposium, celebrating 20 years of atmo-spheric research fostered by Network observations, expands on the Symposium held in Arcachon in 2001 that marked 10 years of measurements. It will coincide with the opening of the NDACC high-altitude station at Maïdo in the Southern Tropics. The aim of the Symposium is to provide a forum to exchange information on the latest scientific achievements using NDACC and related observations, and is structured along five themes:• Long-term evolution and trends in ozone, atmospheric com-

position, temperature, aerosols, and surface UV in the polar regions and at mid-latitudes

• Tropical and sub-tropical observations and analyses• Interactions between atmospheric composition and climate,

in collaboration with NDACC Cooperating Networks• Satellite calibration / validation• New observational capabilities

The Observatoire de Physique de l’Atmosphère de la Réunion (OPAR, http://opar.univ-reunion.fr) is organising the Sympo-sium, which will be held 7-10 November 2011 in Saint Paul, Reunion Island (http://www.reunion.fr). Attendees will be given the opportunity to visit the Maïdo Observatory, currently under construction, which is scheduled to begin operations in early 2012. A Symposium web site will be available in March 2011 for registration, abstract submission, and booking.

Announcement2011 NDACC Symposium

An International Symposium Celebrating 20 Years of Global Atmospheric Research Enhanced by NDACC/NDSC Observations7-10 November 2011, Reunion Island, France

(January 11, 2011 / 08:57:05)

69967-1newsletter36-v2_New_p26.pdf .1

27

Concordiasi is an international effort lead by Météo-France, and involving several research centres and polar institutes (Ra-bier et al., 2010). The project is organised around several observation campaigns in Antarctica, and its main scientific objec-tives are:• to improve the assimilation in numeri-

cal weather prediction (NWP) models of infrared radiances provided by IASI-like hyper-spectral space-borne sound-ers over icy surfaces;

• to enhance the representation of polar processes in numerical models, and in particular to improve the simulation of precipitation and clouds, so as to better describe the mass budget of ice sheets, as well as to provide observational con-straints on gravity-wave drag param-eterizations used in global circulation models;

• to advance our knowledge of micro-physical and dynamical processes in-volved in stratospheric ozone loss, and

Concordiasi: A Project Dedicated to the Polar Atmosphere

F. Rabier, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (florence.rabier@meteo.fr)A. Hertzog, Ecole Polytechnique, CNRS, France (Albert.Hertzog@lmd.polytechnique.fr)Ph. Cocquerez, Centre National d’Etudes Spatiales, France (philippe.cocquerez@cnes.fr)S. A. Cohn, National Center for Atmospheric Research, USA (cohn@ucar.edu)L. Avalonne, University of Colorado, USA (linnea.avallone@lasp.colorado.edu)T. Deshler, University of Wyoming, USA (deshler@uwyo.edu)J. Haase, Purdue University, USA (jhaase@purdue.edu)B. Brioit, Ecole Polytechnique, CNRS, France (brioit @ lmd.polytechnique.fr)F. Danis, Ecole Polytechnique, CNRS, France (francois.danis@lmd.polytechnique.fr)F. Vial, Ecole Polytechnique, CNRS, France (Francois.Vial@lmd.polytechnique.fr)A. Doerenbecher, Centre National de Recherches Météorologiques, CNRS and Mé-téo-France, France (alex.doerenbecher@meteo.fr)V. Guidard, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (vincent.guidard@meteo.fr)D. Puech, Centre National de Recherches Météorologiques, CNRS and Météo-France, France (dominique.puech@meteo.fr)H. Cole, National Center for Atmospheric Research, USA (cole@ucar.edu)J. Fox, National Center for Atmospheric Research, USA (jfox@ucar.edu)T. Hock, National Center for Atmospheric Research, USA (hock@ucar.edu)D. Parsons, National Center for Atmospheric Research, USA (parsons@ucar.edu)J. VanAndel, National Center for Atmospheric Research, USA (vanandel@ucar.edu)L. Kalnajs, University of Colorado, USA (kalnajs@colorado.edu)C. Genthon, Laboratoire de Glaciologie et Géologie de l’Environnement, France (christophe.genthon@lgge.obs.ujf-grenoble.fr)

to better understand the interactions be-tween them.

To this end, a campaign of intensive radio-soundings and surface measurements took place in 2008 and 2009 at Concordia, the French-Italian station on the Antarctic plateau. Radio-soundings were performed at Concordia, and coordinated with the passage of the MetOp satellite (carrying IASI) over the station. In addition, spe-cific ground-based instrumentation was de-ployed at the station, for example snowfall and snow accumulation instruments on the ground, and a 45-m tower equipped with meteorological sensors (wind, moisture, temperature) at several levels to monitor the structure of the polar boundary layer.

The second part of the project is a long-duration balloon campaign that is currently taking place above Antarctica, and will last till early 2011. Nineteen 12-m diameter superpressure balloons were released in

September and October in the stratospheric polar vortex from McMurdo station by the French space agency (CNES). The bal-loons fly at approximately 17 km altitude and carry up to 60 kg of instrumentation and flight devices. Similar balloons have already been successfully used during the Vorcore campaign in 2005, and can typi-cally perform flights that last for several months (Hertzog et al., 2007). Most of the balloons will be flying simultaneously for a few months in the austral spring and early summer, and will provide continuous ob-servations of the polar atmosphere during that period.

All balloons carry a small in situ meteoro-logical package that measures temperature and pressure every 30 seconds. The hori-zontal wind at the flight level is obtained from successive GPS positions of the bal-loons. These observations are sent in near real time to the Global Telecommunication System, so as to be assimilated in the NWP

(January 11, 2011 / 08:57:07)

69967-1newsletter36-v2_New_p27.pdf .1

28

systems operated by the various meteoro-logical services around the world, and thus contribute to the improvement of meteoro-logical forecasts.

Furthermore, thirteen balloons carry the driftsonde gondola developed at NCAR. Each driftsonde gondola contains about 50 miniaturized dropsondes, which can be released individually on demand during the stratospheric balloon flight to provide

high-resolution profiles of thermodynamic variables below the balloon. During the campaign, the drop-soundes are usually coor-dinated with the passage of the MetOp over the balloon location, in order to provide an in situ truth that can be compared with the tem-perature profile retrieved from IASI observations. Some dropsondes are also deployed in the “sensitive regions” of numerical fore-casts, where small improve-ments in the description of the atmospheric flow can lead to large improvements in the simulation.

The remaining 6 balloons carry scientific payloads devoted to tackling scien-tific issues linked to strato-spheric dynamics and chemistry. This equipment provides in situ observa-tion of ozone (with two instruments developed at LMD, France and at the University of Colorado), as well as condensation

nuclei with a particle counter developed by the University of Wyoming. These in situ observations, performed by instruments onboard quasi-Lagrangian tracers, enable us to follow the depletion of ozone during the spring season, and to assess the poten-tial effect of mesoscale waves in triggering the formation of polar stratospheric clouds. In particular, the role of waves generated above the Antarctic Peninsula, which seems

to be important for the formation of PSCs leeward of the mountains, will be moni-tored during the campaign. Furthermore, two of the balloons carry a GPS radio-oc-cultation system developed by the Univer-sity of Purdue to retrieve the temperature profile below the balloon several times per day. These observations, together with the in situ meteorological measurements, will be used to diagnose the stratospheric wave activity over Antarctica. A full report on the preliminary results of the field campaign will be provided in a few months.

Acknowledgements

Concordiasi is an international project, cur-rently supported by the following agencies: Météo-France, CNES, CNRS/INSU, NSF, NCAR, University of Wyoming, Purdue University, University of Colorado, the Al-fred Wegener Institute, the Met Office and ECMWF. Concordiasi also benefits from logistic and/or financial support from the operational polar agencies IPEV, PNRA, USAP and BAS, and from BSRN measure-ments at Concordia. Concordiasi is part of the THORPEX-IPY cluster within the In-ternational Polar Year effort.

References

Hertzog A., et al., 2007: Stratéole-Vorcore - Long-duration, superpressure balloons to study the Antarctic lower stratosphere dur-ing the 2005 winter, J. Atmos. Ocean Tech., 24, 2048-2061.

Rabier F., et al., 2010: The Concordiasi proj-ect in Antarctica, Bull. Am. Meteorol. Soc., 91, 69-86, doi: 10.1175/2009BAMS2764.1.

Balloon launch. Photo courtesy of CNES (Philippe Coc-querez).

The High-Energy-Particle Precipitation in the Atmosphere (HEPPA) Model vs. Data Inter-comparison:

Lessons Learned and Future Prospects

B. Funke, Instituto de Astrofísica de Andalucía, CSIC, Granada, Spain (bernd@iaa.es)

Energetic particle precipitation (EPP) has important implications for atmospheric chemistry. The principal mechanism of EPP affecting the atmospheric composition is the formation of odd nitrogen and hydro-gen radicals, which are both involved in catalytic ozone destruction, via a cascade of dissociation, ionization, and recombina-

tion processes. Solar eruptions and asso-ciated coronal mass ejections sporadically generate intense particle fluxes of very high energy with the potential to penetrate deep into the Earth’s atmosphere in the polar cap regions. During such solar proton events (SPEs), which are more frequent near solar maximum, stratospheric and mesospheric

chemistry can be dramatically altered.

During the last solar cycle, there were sev-eral large SPEs that were intensively ob-served by several instruments on different satellite platforms, including NOAA 16 SBUV/2 and HALOE data (Jackman et al., 2001; Jackman et al., 2005a,b; Randall et

(January 11, 2011 / 08:57:09)

69967-1newsletter36-v2_New_p28.pdf .1

29

The High-Energy-Particle Precipitation in the Atmosphere (HEPPA) Model vs. Data Inter-comparison:

Lessons Learned and Future Prospects

B. Funke, Instituto de Astrofísica de Andalucía, CSIC, Granada, Spain (bernd@iaa.es

chemistry can be dramatically altered.

During the last solar cycle, there were several large SPEs that were intensively ob

SBUV/2 and HALOE data (Jackman et al2001; Jackman et al., 2005a,b; Randall

al., 2005); GOMOS, MIPAS, and SCIA-MACHY on Envisat (Seppälä et al., 2004; López-Puertas et al., 2005a,b; von Clar-mann et al., 2005; Funke et al., 2008; Or-solini et al., 2005; Rohen et al., 2005); and MLS on AURA (Verronen et al., 2006). In particular, during late October and early November 2003, three active solar regions produced solar flares and energetic particles of extremely large intensity; the fourth larg-est event observed in the past forty years. During and after this SPE, often referred as the “Halloween” event, stratospheric and lower mesospheric composition changes were observed in both hemispheres. This includes enormous enhancements in NOx, large depletions in O3, as well as significant changes in other NOy species and N2O. In addition, perturbations of HOx and chlorine species (ClO and HOCl) abundances were observed.

Additionally, magnetospheric electrons precipitate into the atmosphere during geo-magnetic perturbations and generate large amounts of nitric oxide in the polar up-per mesosphere and lower thermosphere. This perturbation occurs throughout the solar cycle, with a maximum intensity ap-proximately 2 years after solar maximum, when the solar wind accelerates. In the ab-sence of sunlight during polar winter, large amounts of EPP-generated odd nitrogen can be transported down to the stratosphere by the meridional circulation without being photochemically destroyed. This mecha-nism is often called the EPP indirect effect. Observational evidence for the EPP-related NOx deposition into the stratosphere has been given by a number of authors (e.g., Siskind et al., 2000; Randall et al., 1998).

Funke et al. (2005) deduced from MIPAS data that a total amount of 2.4 Gmole of NOx was released into the stratosphere dur-ing the Antarctic winter 2003, making up 9% of the stratospheric production due to N2O oxidation in the SH.

EPP indirect effects are strongly linked to the dynamical conditions, showing more pronounced variability in NH polar win-ters than in the SH. Indeed, Randall et al. (2007) have shown that interannual varia-tions of the NOx enhancements in the SH polar winter stratosphere are closely linked to variations of the geomagnetic Ap index (a measure of geomagnetic activity over the globe), suggesting that downward transport of NOx is predominantly controlled by the upper atmospheric EPP source rather than dynamical transport. On the other hand, exceptional dynamical conditions during 3 out of 7 recent NH winters has led to sur-prisingly strong EPP indirect effects there (Randall et al., 2009). In the 2008/2009 winter, large amounts of NOx entered the stratosphere despite low geomagnetic ac-tivity, which clearly demonstrates the im-portance of dynamical modulations of EPP indirect effects in the NH.

Therefore, EPP represents an important solar-terrestrial coupling mechanism that is directly linked to solar variability. The influence of EPP on climate through strato-spheric chemical and dynamical processes is barely understood. This influence is like-ly to extend beyond the polar middle at-mosphere. Evidence for EPP-induced vari-ability in the polar troposphere and tropical stratosphere has recently been demonstrat-ed (i.e., Seppälä et al., 2009, Semeniuk et

al., 2010). A joint effort of both the atmo-spheric modelling and the satellite obser-vation communities is required to advance towards a comprehensive understanding of climate implications, and – in a second step – towards an accurate representation of these effects in climate modelling.

The High Energy Particle Precipitation in the Atmosphere (HEPPA) model vs. data inter-comparison initiative was established during the first HEPPA workshop held in Helsinki in May 2008. It brings together scientists involved in atmospheric model-ling using state-of-the art CCMs and CTMs on one hand, and scientists involved in the analysis and generation of satellite data on the other hand. The objective of this com-munity effort is (i) to assess the ability of state-of-the-art atmospheric models to re-produce EPP-induced composition chang-es, (ii) to identify and, if possible, remedy model deficiencies related to chemistry, dynamics, and ionization schemes, and (iii) to serve as a platform for discussion between modellers and data providers. This is achieved by a quantitative comparison of observed and modelled species abundances during selected periods of pronounced par-ticle forcing, as well as by comparing the simulations performed by different models. In this sense, there is a strong link between the HEPPA model vs. data inter-compari-son initiative and the SOLAR Influence for SPARC (SOLARIS) working group (see article on SOLARIS activities in this is-sue). Both initiatives focus on modelling and understanding the solar influence on climate through chemical and dynamical processes in the middle atmosphere, and complement each other by investigating

Figure 1a: Area-conserving averages (70–90°N) of observed and modelled relative O3 changes with respect to 26 October during 16– 26 November. Thick solid and dashed lines represent multi-model mean average and MIPAS observations, respectively. Figure 1b: Area-conserving averages (40–90°N) of observed and modelled NOy enhancements during 30 October – 1 November with respect to 26 October (left), and relative deviations of modelled averages from the MIPAS observations (right). Thick solid and dashed lines represent multi-model mean average and MIPAS observations, respectively.

-40 -20 0 20 40Ozone change (%)

10

1

Pre

ssur

e (h

Pa)

0.001 0.010 0.100NOy increase (ppmv)

10.0

1.0

0.1

Pre

ssur

e (h

Pa)

-100 -50 0 50 100Difference wrt MIPAS (%)

10.0

1.0

0.1

WACCM

HAMMONIAKASIMA

B3dCTM

EMAC

FinROSE

SOCOLSOCOLi

B2dCTM

CAO

WACCM

HAMMONIAKASIMA

B3dCTM

EMAC

FinROSE

SOCOLSOCOLi

B2dCTM

(a) (b)

(January 11, 2011 / 08:57:10)

69967-1newsletter36-v2_New_p29.pdf .1

30

different aspects of solar variability (par-ticles vs. radiation). See Table 1 for a list of involved scientists.

Past activities

During the last 2 years, the HEPPA initia-tive has focused on the inter-comparison of MIPAS/Envisat data obtained in the after-math of the “Halloween” SPE (26 October – 30 November 2003) with model results. We have compared observations obtained at 25–0.01 hPa in the NH (40-90°N) with simulations performed with the following CCMs and CTMs: the Bremen 2D and 3D Chemical Transport Models (Sinnhuber et al., 2003), FinROSE (Damski et al., 2007), HAMMONIA (Schmidt et al., 2006), KASIMA (Kouker et al., 1999), EMAC (Jöckel et al., 2006), SOCOL and SOCOLi (Egorova et al., 2010), the CAO model (Krivolutsky et al., 2006), and WACCM4 (Garcia et al., 2007). The large number of models participating in the inter-compari-son exercise allowed for an evaluation of the overall ability of atmospheric models to reproduce observed atmospheric per-turbations generated by SPEs, particularly with respect to NOy and ozone changes. This model validation represents a neces-sary first step towards an accurate imple-mentation of particle precipitation effects in long-term climate simulations. Further, the quasi-instantaneous perturbation of the atmosphere due to an SPE acts as a natural laboratory for studying stratospheric and

mesospheric chemistry. This has allowed us to test and to identify deficiencies in the chemical schemes, particularly with respect to nitrogen and chlorine chemistry, both of which are relevant to stratospheric ozone chemistry.

Among the species affected by SPEs, we focused on NO, NO2, N2O, N2O5, HNO3, HNO4, H2O2, O3, ClO, HOCl, and ClONO2. We have further assessed the meteorologi-cal conditions in both the models and the real atmosphere, as observed by MIPAS, by comparing temperature and tracer fields (CH4 and CO). In general, atmospheric models are able to reproduce most of the observed composition changes. In particu-lar, simulated SPE-induced ozone losses agree on average within 5% with the ob-servations. This excellent agreement is found on a short-term scale (HOx-driven) in the mesosphere, as well as on a mid-term scale (NOx-driven) in the stratosphere (see Figure 1a). Simulated NOy enhancements around 1 hPa, however, are on average 30% higher than indicated by the obser-vations (see Figure 1b). This systematic behaviour suggests that these differences are related to the simulated ionization rate profile shape.

The analysis of the observed and modelled NOy partitioning in the aftermath of the “Halloween” SPE has clearly demonstrated the need to implement additional ion chem-istry (e.g., ion-ion recombination between NO3− and H+ cluster ions, and HNO3

formation via water cluster ions) into the chemical schemes. An overestimation of observed H2O2 enhancements by all models hints at an under-estimation of the OH/HO2 ratio in the upper polar stratosphere during the SPE. The analysis of perturbations of the chlorine species ClO, HOCl and ClO-NO2 has shown that the encountered dif-ferences between models and observations, particularly the under-estimation of ob-served ClONO2 enhancements, are related to a smaller availability of ClO in the polar region before the SPE.

In general, the inter-comparison has dem-onstrated that differences in the meteorol-ogy and/or initial state of the atmosphere in the simulations causes a significant vari-ability of the model results, even on the short time scale of only a few days. Fur-thermore, this sensitivity of the simulated atmospheric responses to the background conditions, indicated by the spread in the model results, also implies that the response to proton events in the real atmosphere de-pends strongly on the actual conditions.

Future activities

At the second HEPPA workshop held in Boulder, CO in October 2009, it was de-cided to focus future activities on EPP indi-rect effects (i.e., polar winter NOx descent). This decision was motivated by the higher potential of EPP indirect effects to influ-ence middle atmospheric composition on longer time scales compared to direct ef-fects (i.e., SPEs), and by the large variabili-ty in EPP indirect effects related to dynami-cal modulations, making its representation in current atmospheric models challenging.

In particular, the 2008/2009 NH polar win-ter turned out to be a very interesting period because of the peculiar dynamic conditions which were characterized by an unusually strong and persistent stratospheric sudden warming (SSW) that occurred in January, followed by the reformation of a strong up-per stratospheric vortex, with very efficient descent of the vortex during the following weeks. MIPAS observations of NOx and the tracer CO in the 70-90°N region (see Figure 2 - colour plate III) illustrate the dynamical modulations of the EPP indirect effects dur-ing this particular NH winter: After the SSW in late January, which provoked a depletion of CO and NOx in the upper stratosphere and mesosphere, large amounts of these spe-cies descended very quickly into the upper

A. Baumgärtner abaumg@mpch-mainz-mpg.de MPI, Mainz, GermanyM. Calisto marco.calisto@env.ethz.ch IAC/ETH, Zürich, SwitzerlandT. Egorova t.egorova@pmodwrc.ch PMOD/WRC, Davos,

SwitzerlandB. Funke bernd@iaa.es IAA, Granada, SpainC. H. Jackman Charles.H.Jackman@nasa.gov NASA-GFSC, MD, USAM.-B. Kallenrode mkallenr@uos.de Univ. of Osnabrück,

Germany J. Kieser jens.kieser@zmaw.de MPI-M, Hamburg, GermanyA. Krivolutsky alexei.krivolutsky@rambler.ru CAO, Moscow, RussiaM. López-Puertas puertas@iaa.es IAA, Granada, SpainD. Marsh marsh@ucar.edu NCAR, Boulder, CO, USAC. Randall Cora.Randall@lasp.colorado.edu LASP, Boulder, CO, USAT. Reddmann thomas.reddmann@kit.edu KIT, Karlsruhe, GermanyE. Rozanov e.rozanov@pmodwrc.ch PMOD/WRC and IAC/ETH,

Davos, SwitzerlandS.-M. Salmi sanna-mari.salmi@fmi.fi FMI, Helsinki, FinlandM. Sinnhuber miriam.sinnhuber@kit.edu KIT, Karlsruhe, GermanyH. Schmidt hauke.schmidt@zmaw.de MPI-M, Hamburg, GermanyG. P. Stiller gabriele.stiller@kit.edu KIT, Karlsruhe, GermanyP. T. Verronen pekka.verronen@fmi.fi FMI, Helsinki, FinlandS. Versick stefan.versick@kit.edu KIT, Karlsruhe, GermanyT. von Clarmann thomas.clarmann@kit.edu KIT, Karlsruhe, GermanyN. Wieters nwieters@iup.physik.uni-bremen.de Univ. of Bremen, GermanyJ. M. Wissing jawissin@uos.de Univ. of Osnabrück, Germany

Table 1: Involved Scientists

(January 11, 2011 / 08:57:12)

69967-1newsletter36-v2_New_p30.pdf .1

31

stratosphere. Similar situations were also found in the NH during January 2004 and 2006, however, the period November 2008 – May 2009 was better covered by satellite data than the previous winters.

The planned inter-comparison exercise will focus on the assessment of the EPP source (and its spatial distribution) by analysing observed and modelled NOx distributions from the mesosphere up to the thermo-sphere (60-150 km) during the period of interest; the analysis of vertical coupling mechanisms by inter-comparison of ob-served and modelled tracer and tempera-ture fields with particular emphasis on the MLT region; and the assessment of strato-spheric mid-term composition changes in-duced by EPP indirect effects during 2009 with particular emphasis on ozone and NOy repartitioning.

Spatially resolved observational data dur-ing November 2008 – May 2009 is avail-able from a large number of instruments (e.g., ACE-FTS; GOMOS, MIPAS, and SCIAMACHY on Envisat; MLS/Aura; SMR/Odin; SABER/TIMED), including NO and temperature up to the middle ther-mosphere, O3 and CO up to 100 km, as well as stratospheric and mesospheric distribu-tions of NO2, HNO3, N2O5, N2O, and CH4.

Interested scientists working on satellite data or atmospheric modelling are encour-aged to participate in this activity. Atmo-spheric models to be included in the inter-comparison should have the capacity to be nudged to meteorological analyses during the period of interest, and preferably cover the MLT region. It has been shown, how-ever, that models with lower lids can suc-cessfully be employed for simulations of EPP indirect effects by means of adequate-ly chosen upper boundary conditions (Vo-gel et al., 2008, Reddmann et al., 2010). The expected outcome of this new inter-comparison exercise is (i) the validation of EPP implementations in atmospheric models, (ii) a better understanding of the EPP source distribution and vertical cou-pling mechanisms, and (iii) the quantifica-tion (and model validation) of EPP indirect effects on stratospheric chemistry. A dedi-cated workshop will be held during the up-coming third HEPPA meeting in Granada/Spain (9-11 May 2011, http://heppa2011.iaa.es/), focusing on the coordination of these activities and providing an opportu-nity to present first results.

References

Damski, J., et al., 2007: FinROSE - middle at-mospheric chemistry transport model, Boreal Env. Res., 12, 535-550.

Egorova, T., et al., 2010: The atmospheric effects of October 2003 solar proton event simulated with the chemistry–climate model SOCOL using complete and parameterized ion chemistry, J. Atmos. Sol. Terr. Phys., in press, doi:10.1016/j.jastp.2010.01.009.

Funke, B., et al., 2005: Downward transport of upper atmospheric NOx into the polar strato-sphere and lower mesosphere during the Antarc-tic 2003 and Arctic 2002/2003 winters, J. Geo-phys. Res., 110, doi:10.1029/2005JD006463.

Funke, B., et al., 2008: Enhancement of N2O during the October–November 2003 solar pro-ton events, Atmos. Chem. Phys., 8, 3805–3815.

Garcia, R. R., et al., 2007: Simulation of secular trends in the middle atmosphere, 1950-2003, J. Geophys. Res., 112, doi:10.1029/ 2006JD007485.

Jackman, C. H., et al., 2001: Northern hemisphere atmospheric effects due to the July 2000 solar pro-ton event, Geophys. Res. Lett., 28, 2883–2886.

Jackman, C. H., et al., 2005a: Neutral atmo-spheric influences of the solar proton events in October–November 2003, J. Geophys. Res., 110, doi:10.1029/ 2004JA010888.

Jackman, C. H., et al., 2005b: The influence of the several very large solar proton events in Years 2000–2003 on the neutral middle atmo-sphere, Adv. Space Res., 35, 445–450.

Jöckel, P., et al., 2006: The atmospheric chem-istry general circulation model ECHAM5/MESSy1: consistent simulation of ozone from the surface to the mesosphere, Atmos. Chem. Phys., 6, 5067-5104.

Kouker, W., et al., 1999: Streamers observed by the CRISTA experiment and simulated in the KASI-MA model, J. Geophys. Res., 104, 16,405-16,418.

Krivolutsky et al., 2006: Dynamical response of the middle atmosphere to solar proton event of July 2000: Three-dimensional model simula-tions, Adv. Space Res., 37, 1602-1613.

López-Puertas, M., et al., 2005a: Observa-tion of NOx enhancement and ozone deple-tion in the Northern and Southern Hemi-spheres after the October–November 2003 solar proton events, J. Geophys. Res., 110, doi:10.1029/2005JA011050.

López-Puertas, M., et al., 2005b: HNO3, N2O5, and ClONO2 enhancements after the October–November 2003 solar proton events, J. Geo-phys. Res., 110, doi:10.1029/2005JA011051.

Orsolini, Y. J., et al., 2005: An upper strato-spheric layer of enhanced HNO3 following ex-

ceptional solar storms, Geophys. Res. Lett., 32, doi:10.1029/2004GL021588.

Randall, C. E., et al., 1998: Polar ozone and aero-sol measurement (POAM) II stratospheric NO2, 1993–1996, J. Geophys. Res., 103, 28,361–28,371.

Randall, C. E., et al., 2007: Energetic particle precipitation effects on the Southern Hemi-sphere stratosphere in 1992–2005, J. Geophys. Res., 112, doi:10.1029/2006JD007696.

Randall, C. E., et al., 2009: NOx descent in the Arctic middle atmosphere in early 2009, Geo-phys. Res. Lett., 36, doi:10.1029/2009GL039706.

Reddmann, T., et al., 2010: Modeling dis-turbed stratospheric chemistry during solar-induced NOx enhancements observed with MIPAS/ENVISAT, J. Geophys. Res., 115, doi:10.1029/2009JD012569.

Schmidt H., et al., 2006: The HAMMONIA chemistry climate model: Sensitivity of the me-sopause region to the 11-year solar cycle and CO2 doubling, J. Clim., 19, 3903-3931.

Semeniuk, K., et al., 2010: Middle atmosphere response to the solar cycle in irradiance and ion-izing particle precipitation, Atmos. Chem. Phys. Discus., 10, 24,853-24,917.

Seppälä, A., et al., 2004: Solar proton events of October–November 2003: Ozone depletion in the Northern Hemisphere polar winter as seen by GOMOS/Envisat, Geophys. Res. Lett., 31, doi:10.1029/2004GL021042.

Seppälä, A., et al., 2009: Geomagnetic activity and polar surface air temperature variability, J. Geophys. Res., 114, doi:10.1029/2008JA014029.

Sinnhuber, M., et al., 2003: A model study of the impact of magnetic field structure on atmo-spheric composition during solar proton events. Geophys. Res. Lett., 30, 1818-1821.

Siskind, D. E., 2000: On the coupling between middle and upper atmospheric odd nitrogen, in Atmospheric Science Across the Stratopause, Geophys. Monogr. Ser., vol. 123, edited by D. E. Siskind, S. D. Eckermann, and M. E. Sum-mers, pp. 101–116, AGU, Washington, D. C.

Verronen, P. T., et al., 2006: Production of odd hydrogen in the mesosphere during the January 2005 solar proton event, Geophys. Res. Lett., 33, doi:10.1029/2006GL028115.

Vogel, B., et al., 2008: Model simulations of stratospheric ozone loss caused by enhanced me-sospheric NOx during Arctic Winter 2003/2004, Atmos. Chem. Phys., 8, 5279.

von Clarmann, T., et al., 2005: Stratospheric HOCl measurements provide evidence of in-creased stratospheric odd hydrogen abundances due to solar proton events 2003, J. Geophys. Res., 110, doi:10.1029/2005JA011053.

(January 11, 2011 / 08:57:14)

69967-1newsletter36-v2_New_p31.pdf .1

32

Stratospheric Change and its Role for Climate Prediction (SHARP): A contribution to SPARC

U. Langematz, Freie Universität Berlin, Germany (ulrike.langematz@met.fu-berlin.de) and the SHARP consortium

Since June 2009, scientists and students from eight German institutions have been working together on SPARC-related sci-ence issues in the research unit Stratospher-ic Change and its Role for Climate Predic-tion (SHARP) funded by the Deutsche Forschungsgemeinschaft (DFG; German Science Foundation). The proposal for SHARP was strongly motivated by the New SPARC Initiatives and this article aims to introduce the goals and current research activities of SHARP to the international SPARC community. International part-ners (e.g., Bodeker Scientific, University Cambridge, UK Met Office, University of Utrecht, and Columbia University in New York) are associated members of SHARP.

The primary objective of SHARP is to im-prove our understanding of global climate change and the accuracy of climate change predictions, with emphasis on the relevance of the stratosphere. SHARP is coordinating research activities in Germany with two leading themes:• The interactions between climate

change, stratospheric dynamics and at-mospheric composition.

• The interaction between stratospheric change and tropospheric climate and weather.

To foster optimal collaboration between modelling and measurement groups and across the locally distributed institutions, four collaborative scientific projects have been defined in SHARP that address the following current key research aspects:1. The detection, investigation and ex-

planation of recent and potential future changes in the Brewer-Dobson circu-lation and their implications for strato-spheric dynamics, physics and chemis-try in a changing climate. This combines optimised retrievals of atmospheric data products and simulations of improved Chemistry Climate Models (CCMs) and General Circulation Models (GCMs).

2. The detection and attribution of changes in stratospheric ozone (O3) during the anticipated turnaround of chlorine load-ing, and the prediction of O3 change in response to and as a result of feedback

with global climate change.

3. The explanation of recent stratospheric water vapour (H2O) concentration changes by extending the time series of ground based and satellite data prod-ucts in conjunction with model studies, and a reliable assess-ment of future H2O concentrations based on the improved un-derstanding about the key processes gained from studying the past.

4. The attribution and prediction of chang-es in tropospheric weather and climate in response to stratosphere-troposphere coupling, and our understanding of the underlying mechanisms based on atmo-spheric observations and simulations of CCMs and GCMs.

To achieve these goals, leading German modelling and measurement research groups have organised and coordinated their research in a synergistic and com-plmentary effort. SHARP makes use of:• Measurements of stratospheric com-

position, in particular from the SCIAMACHY satellite instrument of University Bremen and the MIPAS instrument of Karlsruhe Institute for Technology, for the analysis of strato-spheric change and the validation of the model simulations. For the derivation of long-term trends, the data analysis is supported by measurements from balloon platforms of the Universities Frankfurt and Heidelberg. In addition the SHARP team collaborates with the German Weather Service (DWD) long term measurement programme.

• The EMAC-FUB and E39C-A Chem-istry-Climate Models (CCMs) run at Freie Universität Berlin (FUB) and DLR, which simulate the complex in-teractions between chemical processes, dynamics and radiative forcing for the

attribution and prediction of climate change. The CCM studies are supported by sensitivity studies with the ECHAM5 General Circulation Models (GCMs) of MPI for Meteorology (MPIM) and the ECHAM5 and EGMAM Atmosphere-Ocean GCMs (AOGCM) of MPIM and FUB to investigate natural variability and separate the effects of specified cli-mate forcings.

Table 1 gives a summary of the SHARP consortium and the contributions of the individual members to the research unit. Currently, one post-doc, 8 PhD students, 4 student assistants, and one administrative assistant are employed in SHARP proj-ects. The photo shows the SHARP group at the first annual meeting in Bremen in May 2010. The research unit is coordinated at Freie Universität Berlin. More informa-tion can be found on the SHARP website www.fu-berlin.de/sharp/. The following sections present an overview of the objec-tives of the four individual science projects and selected new results.

Project SHARP-BDC

In the project SHARP-BDC, coordinated by Martin Dameris (DLR), the most im-portant focus addressed is “How is the Brewer-Dobson circulation affected by climate change, and which processes are relevant?” In this project, dynami-cal, physical and chemical processes, as

SHARP staff at 2010 annual meeting in Bremen.

well as feedback effects relevant for the stratospheric residual circulation (Brewer-Dobson circulation, BDC) are investigated. Moreover, the impact of changes in atmo-spheric composition and climate on these processes are studied in detail using nu-merical simulations with the Chemistry-Climate Models (CCMs) EMAC-FUB and E39C-A in connection with observations. The influences of atmospheric changes on the BDC will be identified and quantified, as well as their feedback on tracer distributions and surface climate (see also SHARP-STC).

First results of the climatology and trends in tropical upwelling in the lower strato-sphere simulated with E39C-A have been presented in Stenke et al. (2009) and Gar-ny et al. (2009). The aim was to quantify changes in tropical upwelling and examine potential contributing mechanisms. The drivers of upwelling in the tropical lower stratosphere were investigated using re-sults of different multi-decadal simulations (transient and in time-slice mode) of E39C-A. The climatological annual cycle in up-welling and its wave forcing were validated against ERA-Interim analysis. It turned out that the strength in tropical upwelling and its annual cycle can be largely explained by

Participants Institution/Institute Contribution to SHARP Scientific Area Ulrike Langematz Freie Universität Berlin (FUB), Institut für

MeteorologieSpeaker of SHARP, PI of SHARP-STC, Co-I of SHARP-BDC, SHARP-OCF, SHARP-WV

Chemistry-Climate Modelling, EMAC-FUB CCM

Martin Dameris Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Atmosphärenphysik

PI of SHARP-BDC, Co-I of SHARP-OCF, SHARP-WV, SHARP-STC

Chemistry-Climate Modelling, E39C-A CCM

John P. Burrows Universität Bremen (UBR), Institut für Umweltphysik

PI of SHARP-OCF Stratospheric trace gases, GOME, SCIAMACHY

Gabriele Stiller Karlsruher Institut für Technologie (KIT), Institut für Meteorologie und Klimaforschung

PI of SHARP-WV, Co-I of SHARP-BDC, SHARP-OCF

Stratospheric trace gases, MIPAS

Christoph Brühl Max-Planck-Institut für Chemie (MPIC) Co-I of SHARP-OCF Chemistry-Climate Modelling,EMAC CCM

Ulrich Cubasch Freie Universität Berlin (FUB), Institut für Meteorologie

Co-I of SHARP-STC Climate Modelling, AO-GCM EGMAM

Andreas Engel Goethe Universität Frankfurt (JWGU), Institut für Atmosphäre und Umwelt

Co-I of SHARP-BDC, SHARP OCF Stratospheric trace gases,Balloon-borne whole air sampler

Marco Giorgetta Max-Planck-Institut für Meteorologie (MPIM) Co-I of SHARP-BDC, SHARP WV, SHARP- STC

Climate modelling, ECHAM GCM and AO-GCM

Patrick Jöckel Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Atmosphärenphysik

Co-I of SHARP-WV Water isotopes, EMAC CCM

Klaus Pfeilsticker Universität Heidelberg (UH), Institut für Umweltphysik

Co-I of SHARP-OCF Stratospheric trace gases, LPMA/DOAS balloon measurements

Björn-Martin Sinnhuber

Karlsruher Institut für Technologie (KIT), Institut für Meteorologie und Klimaforschung

Co-I of SHARP-OCF VSLS, CTM-Modelling

Mark Weber Universität Bremen (UBR), Institut für Umweltphysik

Co-I of SHARP-OCF, SHARP-WV Trace gases and dynamics, SCIAMACHY, GOME

Table 1: Principle Investigators (PI) and Co-Investigators (Co-I) in the SHARP research unit.

(January 11, 2011 / 08:57:16)

69967-1newsletter36-v2_New_p32.pdf .1

33

well as feedback effects relevant for the stratospheric residual circulation (Brewer-Dobson circulation, BDC) are investigated. Moreover, the impact of changes in atmo-spheric composition and climate on these processes are studied in detail using nu-merical simulations with the Chemistry-Climate Models (CCMs) EMAC-FUB and E39C-A in connection with observations. The influences of atmospheric changes on the BDC will be identified and quantified, as well as their feedback on tracer distributions and surface climate (see also SHARP-STC).

First results of the climatology and trends in tropical upwelling in the lower strato-sphere simulated with E39C-A have been presented in Stenke et al. (2009) and Gar-ny et al. (2009). The aim was to quantify changes in tropical upwelling and examine potential contributing mechanisms. The drivers of upwelling in the tropical lower stratosphere were investigated using re-sults of different multi-decadal simulations (transient and in time-slice mode) of E39C-A. The climatological annual cycle in up-welling and its wave forcing were validated against ERA-Interim analysis. It turned out that the strength in tropical upwelling and its annual cycle can be largely explained by

local, resolved wave forcing. The climato-logical mean forcing is due to both station-ary planetary-scale waves that originate in the trop-ics, and to extra-tropical transient synoptic scale waves that are refracted equatorward. In the CCM, further increases in atmo-spheric greenhouse gas concentrations to the year 2050 force a year around positive trend in tropical upwelling, maximising in the lowermost strato-sphere. Tropical ascent is balanced by downwelling between 20° and 40°. In-creases in tropical upwell-ing can be explained by stronger local forcing by resolved wave conver-gence, which is driven in turn by processes initiated by increases in tropical sea surface temperatures (SSTs). Higher tropical SSTs cause a strengthen-ing of the subtropical jets and modification of deep

convection affecting latent heat release. While the former can modify wave propa-

Participants Institution/Institute Contribution to SHARP Scientific Area Ulrike Langematz Freie Universität Berlin (FUB), Institut für

MeteorologieSpeaker of SHARP, PI of SHARP-STC, Co-I of SHARP-BDC, SHARP-OCF, SHARP-WV

Chemistry-Climate Modelling, EMAC-FUB CCM

Martin Dameris Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Atmosphärenphysik

PI of SHARP-BDC, Co-I of SHARP-OCF, SHARP-WV, SHARP-STC

Chemistry-Climate Modelling, E39C-A CCM

John P. Burrows Universität Bremen (UBR), Institut für Umweltphysik

PI of SHARP-OCF Stratospheric trace gases, GOME, SCIAMACHY

Gabriele Stiller Karlsruher Institut für Technologie (KIT), Institut für Meteorologie und Klimaforschung

PI of SHARP-WV, Co-I of SHARP-BDC, SHARP-OCF

Stratospheric trace gases, MIPAS

Christoph Brühl Max-Planck-Institut für Chemie (MPIC) Co-I of SHARP-OCF Chemistry-Climate Modelling,EMAC CCM

Ulrich Cubasch Freie Universität Berlin (FUB), Institut für Meteorologie

Co-I of SHARP-STC Climate Modelling, AO-GCM EGMAM

Andreas Engel Goethe Universität Frankfurt (JWGU), Institut für Atmosphäre und Umwelt

Co-I of SHARP-BDC, SHARP OCF Stratospheric trace gases,Balloon-borne whole air sampler

Marco Giorgetta Max-Planck-Institut für Meteorologie (MPIM) Co-I of SHARP-BDC, SHARP WV, SHARP- STC

Climate modelling, ECHAM GCM and AO-GCM

Patrick Jöckel Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Atmosphärenphysik

Co-I of SHARP-WV Water isotopes, EMAC CCM

Klaus Pfeilsticker Universität Heidelberg (UH), Institut für Umweltphysik

Co-I of SHARP-OCF Stratospheric trace gases, LPMA/DOAS balloon measurements

Björn-Martin Sinnhuber

Karlsruher Institut für Technologie (KIT), Institut für Meteorologie und Klimaforschung

Co-I of SHARP-OCF VSLS, CTM-Modelling

Mark Weber Universität Bremen (UBR), Institut für Umweltphysik

Co-I of SHARP-OCF, SHARP-WV Trace gases and dynamics, SCIAMACHY, GOME

Figure 1: Schematic of the two branches of the meridional circulation in the stratosphere, and its wave driving. Wave flux convergence is indicated in light grey patches (negative EP divergence). The global classical BDC (a) is driven by extra-tropical waves, and a deep hemisphere-wide cell exists in the winter hemisphere. The secondary circulation (b) is confined to the (sub-) tropical lower stratosphere, and driv-en locally by wave dissipation. Both tropical waves (mostly generated by strong deep convection in the summer tropics) and extra-tropical waves (mostly refracted to low latitudes) contribute to the wave convergence in the upper troposphere/lower stratosphere. Figure taken from Garny et al., 2010.

Table 1: Principle Investigators (PI) and Co-Investigators (Co-I) in the SHARP research unit.

(January 11, 2011 / 08:57:18)

69967-1newsletter36-v2_New_p33.pdf .1

34

gation and dissipation, the latter affects tropical wave generation. The dominant mechanism leading to enhanced vertical wave propagation into the lower strato-sphere is an upward shift of the easterly shear zone due to the strengthening and up-ward and equatorward shift of the subtropi-cal jets. A summary of the mechanisms is given in Figure 1. More details about this study can be found in Garny et al. (2010).

Project SHARP-OCF

SHARP-OCF focuses on the question “How is the evolution of stratospheric ozone affected by climate change, and how strong is the feedback?” SHARP-OCF is coordinated by John P. Burrows, University Bremen. One major goal of this project is to analyse present observa-tional trace gas data together with state-of-the art models in order to obtain a better understanding of the interaction between ozone and climate change and the underly-

ing dynamical and chemical processes. Satellite, balloon and aircraft obser-vations are used to assess the bud-gets and changes/trends of strato-spheric ozone and the key halogenated substances, in par-ticular, very short-lived substances (VSLS). Transient CCM simulations and supplementary sensitivity studies

together with the observational data record are to be analysed to assess past and future evolution of stratospheric ozone and other key species.

Long-term total ozone data sets are now available from the “European” satellites GOME1, SCIAMACHY and GOME2 starting in 1995, and provide both total col-umn and vertical profiles in an early morn-ing orbit. Before these data sets can be used for long-term trend assessments, any biases and possible drifts between instruments must be removed. This has been done by matching the SCIAMACHY and GOME2 data record to GOME1. Using zonal-mean monthly mean data, the drifts and biases for SCIAMACHY and GOME2 have been corrected and a merged data set produced, which is called the GSG merged (GOME1/SCIAMACHY/GOME2) data set (http://www.iup.uni-bremen.de/gome/wfdoas_merged.html, Weber et al., 2007). Figure

2 highlights the interan-nual variability of the GSG data set shown as anomalies with respect to the seasonal average from 1995-2009. The cold Arc-tic winters in the mid-1990s with severe polar ozone losses, the Antarc-tic ozone hole anomaly in 2002, as well as the record ozone hole in 2006 are clearly seen. In both the tropics and extra-trop-ics, the QBO signal is a prominent feature.

Some of the recent work on stratospher-ic BrO retrieved from

SCIAMACHY is shown in Figure 3. The integrated BrO between 2002 and 2010 is plotted for 50°-60°N, 20°N-20°S, and 50°-60°S. Both the seasonal variation at mid-latitudes and the longer term decrease of BrO are clearly observed. Detailed analy-sis will be undertaken within SHARP.

Transient CCM simulations with the E39C-A and EMAC-FUB models that have been performed within SHARP contributed to the CCMVal initiative (SPARC CCMVal, 2010) and were part of the projections of the future evolution of ozone for the up-coming WMO Assessment of Stratospheric Ozone: 2010. Figure 4 (see colour plate IV) shows that most CCMs project that the ozone hole will vanish with respect to their 1960-1965 minimum area in the second half of the 21st century, however with a large uncertainty in the return date (Austin et al., 2010).

Project SHARP-WV

SHARP-WV focuses on stratospheric wa-ter vapour and the question: How is strato-spheric water vapour affected by climate change, and which processes are responsi-ble? SHARP-WV is coordinated by Gabri-ele Stiller (KIT Karlsruhe). SHARP-WV will analyse observational data sets from the satellite instruments MIPAS and SCIA-MACHY, merged with the HALOE and SAGE data sets, and data from long-term simulations with CCMs in order to improve our understanding of past variations and trends in stratospheric H2O, and to assess the future evolution of the stratospheric H2O budget in a changing climate.

In particular, the satellite observations will be used to study the stratospheric water va-pour distribution and its temporal (on vari-ous scales) and spatial anomalies, as well as changes on a decadal scale. The tropical and extra-tropical mechanisms for water vapour transport into the stratosphere (e.g., monsoon activity) and their relative impor-tance will also be investigated, making ad-ditional use of the isotopic composition of stratospheric water vapour, which is provid-ed by MIPAS observations. Series of multi-year simulations with several different set-ups of the CCMs ECHAM5/MESSy and E39C-A will be analysed in the same way as the observational data in order to validate the understanding of relevant processes of transport, and stratospheric sources and sinks under present and future conditions.

Figure 3: The change/trend of BrO at 50°-60°N, 20°N-20°S and 50°-60°S from SCIAMACHY (A. Rozanov and J. P. Bur-rows IUP University of Bremen).

Figure 2: Total ozone anomaly from the merged GOME1/SCIA-MACHY/GOME2 (GSG) data set. The anomalies are calculated with respect to the seasonal mean from 1995-2006 (adapted from Weber and Steinbrecht, 2010).

Total ozone anomaly (reference period 1996-2009)La

titud

e

50N

0

50S

DU

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

-200 -100 -50 -25 -15 -10 -5 5 10 15 25 50 100 200

BrO

col

umn

anom

aly

17-2

7 km

(1012

mol

ecul

e/cm

2 )

2.0

1.5

1.0

0.5

0.0

-0.5

-1.0

-1.5

-2.0

Year2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

50-60ºN (Trend = -0.9%/year)20ºS-20ºN (Trend = -1.8%/year)50-60ºS (Trend = -1.0%/year)

50-60ºN (Trend = -0.9%/year)50 60ºN (T d 0 9%/ )20ºS-20ºN (Trend = -1.8%/year)

(January 11, 2011 / 08:57:20)

69967-1newsletter36-v2_New_p34.pdf .1

35

First results on the analysis of water vapour transport through the Indian monsoon anti-cyclone have been published by Kunze et al. (2010). Figure 5 (see colour plate IV) shows the distribution of water vapour at 360 K in the region of the Asian Monsoon Anticyclone (AMA) as a four-year average of July-August MIPAS observations and a long-term monthly mean for the three CCMs involved in SHARP, respectively. Although the absolute water vapour mixing ratios between observation and models dif-fer, the overall structure of enhanced water vapour, hinting towards upward transport in the AMA, is well reproduced. In detail, however, the models differ considerably regarding the position of the water va-pour maximum relative to the centre of the AMA.

For the first time, vertical distributions of stratospheric water vapour were ob-tained from space borne limb observa-tions of the scattered solar radiation using SCIAMACHY (Rozanov et al., 2010). Within SHARP-WV, it is planned to pro-duce time series of zonal mean water va-pour for the entire SCIAMACHY obser-vation period beginning in August 2002. The water vapour retrieval is fairly time consuming because multiple scattering must be considered, in particular from the troposphere where water vapour is several orders of magnitude more abundant than in the stratosphere. Figure 6 (see colour plate IV) shows the zonal mean water vapour volume mixing ratios derived from SCIA-MACHY using ECMWF temperatures and pressures as input into the retrievals in the

zonal bands 40°N - 45°N (a) and 40°S - 45°S (b) for altitudes between 10 and 25 km. The annual cycle in each hemisphere is clearly visible from these data.

Project SHARP-STC

SHARP-STC deals with stratosphere-troposphere coupling and the question: “How is the coupling of the stratosphere and troposphere affected by climate change, and how strong is the feedback on climate?” The project is coordinated by Ulrike Langematz (FUB). The focus of SHARP-STC is to determine the role of the interaction between the stratosphere and troposphere in a changing climate, in particular to assess the impact of a chang-ing stratosphere on surface climate and weather.

Figure 7 illustrates the dynamical cou-pling between the stratosphere and tro-posphere in five Northern Hemisphere winters of a 300-year simulation with the Atmosphere-Ocean GCM (AO-GCM) EG-MAM (Langematz et al., 2010). Negative anomalies in the signature of the Northern Annular Mode (NAM) that are associated with major warmings, as for example in February 2001, propagate downward into the troposphere, where they modify weath-er patterns with a delay of several weeks. Similarly, positive NAM anomalies associ-ated with intense stratospheric polar vorti-ces are followed by tropospheric positive anomalies.

In SHARP-STC, the transient simulations of the past and future with the EMAC-FUB

and E39C-A CCMs are analysed to study how well current models are able to reproduce the observed stratosphere-troposphere coupling, to understand the respon-sible mechanisms, and to assess its future evo-lution. Complementary sensitivity simulations will be performed with a spectrum of models of different complex-ity (GCMs with dif-ferent horizontal and vertical resolution, with and without coupled ocean and chemistry) to isolate the effects of

changes in greenhouse gases, stratospheric ozone, water vapour and sea surface tem-peratures on near-surface climate through downward coupling.

References

Austin, J., et al., 2010: Chemistry-climate mod-el simulations of spring Antarctic ozone, J. Geo-phys. Res., 115, doi: 10.1029/2009JD013577.

Garny, H., M. Dameris, and A. Stenke, 2009: Impact of prescribed SSTs on climatologies and long-term trends in CCM simulations, Atmos. Chem. Phys., 9, 6017-6031.

Garny, H., et al., 2010: Dynamically forced in-crease of tropical upwelling in the lower strato-sphere, J. Atmos. Sci., submitted.

Kunze, M., et al., 2010: Influences of the Indian Summer Monsoon on water vapor and ozone con-centrations in the UTLS as simulated by Chemis-try-Climate Models, J. Clim., 23, 3525-3544.

Langematz, U., et al., 2010: Stratosphere-tropo-sphere dynamical variability in an atmosphere-ocean general circulation model (AO-GCM), J. Geophys. Res., submitted.

Rozanov, A., et al., 2010: Retrieval of water vapor vertical distributions in the upper tropo-sphere and the lower stratosphere from SCIA-MACHY limb measurements, Atmos. Meas. Tech. Discuss., 3, 4009-4057.

SPARC CCMVal, 2010: Report on the Evalu-ation of Chemistry-Climate Models, V. Eyring, T. G. Shepherd, D. W. Waugh (Eds.), SPARC Report No. 5, WCRP-132, WMO/TD-No. 1526,

Stenke, A., et al., 2009: Implications of La-grangian transport for simulations with a cou-pled chemistry-climate model, Atmos. Chem. Phys., 9, 5489-5504.

Weber, M., et al., 2003: Dynamical control of NH and SH winter/spring total ozone from GOME observations in 1995-2002, Geophys. Res. Lett., 30, doi:10.109/2002GL016799.

Weber, M., L. N. Lamsal, and J. P. Burrows, 2007: Improved SCIAMACHY WFDOAS total ozone retrieval: Steps towards homogenising long-term total ozone datasets from GOME, SCIAMACHY, and GOME2, Proc. ‘Envisat Symposium 2007’, Montreux, Switzerland, 23-27 April 2007, ESA SP-636, July 2007.

Weber, M., and W. Steinbrecht, 2010: Strato-spheric Ozone, [in “State of the Climate in 2009”] Bull. Amer. Meteor. Soc., 91, 46-49.

Figure 7: 90-day low-pass filtered time series of the NAM sig-nature for five model years of a 300-year present day simulation with the EGMAM AOGCM. Positive values represent negative polar geopotential height anomalies and strong stratospheric vortices; negative values represent positive polar geopoten-tial height anomalies and weak stratospheric vortices. From Langematz et al. (2010).

07/2200-06/2205 Time series of the AO index

90-day Low pass filter

1.5

1.0

0.5

-0.5

-1.0

-1.5

Pre

ssur

e (h

Pa)

1JAN01 1JUL 1JAN02 1JUL 1JAN03 1JUL 1JUL1JAN051JUL1JAN04

0.030.05

0.1

0.30.5

1

35

10

3050

100

300500

1000

MAJOR

(January 11, 2011 / 08:57:24)

69967-1newsletter36-v2_New_p35.pdf .1

36

Director: N. McFarlane Project Scientist: D. Pendlebury Administrator: M. Rosen

SPARC IPO Department of Physics University of Toronto, 60 St. George St. Toronto, ON M5S 1A7 - Canada Tel: (1) 416 946 7543 Fax: (1) 416 946 0513 Email: sparc@atmosp.physics.utoronto.ca http://www.atmosp.physics.utoronto.ca/SPARC/

Co-ChairsT. Peter (Switzerland)T.G. Shepherd (Canada) SSG Members (2011)

G. Bodeker (New Zealand)J.P. Burrows (Germany)Hong-Bin Chen (China)P.C.S. Devara (India)R. Diab (South Africa)

V. Eyring (Germany)D. Fahey (USA)M. Pulido (Argentina)M. Santee (USA)A. Scaife (United Kingdom)M. Shiotani (Japan)

Ex-Officio MembersCOSPAR: J. Gille (USA)IGAC: S. Doherty (USA)NDACC: M. Kurylo (USA) SCOSTEP: M. Geller (USA) WMO/GAW: G.O. Braathen (Switzerland)

SPARC Scientific Steering Group

JPS/WCRP: V. Ryabinin (Switzerland)

Composition of the SPARC Office

Liaison with WCRP

201123-27 January 91st Annual Meeting of AMS, Seattle, WA, USA, http://www.ametsoc.org/MEET/meetinfo.html.

- 16th Confrence on Middle Atmosphere - 23rd Conference on Climate Variability and Change

28 February-4 March

AGU Chapman Conference on “Atmospheric Gravity Waves and their Effects on General Circulation and Climate”, Honolulu, Hawaii, http://www.agu.org/meetings/chapman/

5-9 June Canadian Meteorological and Oceanographic Society Congress (CMOS/SCMO)Victoria, BC, Canada http://www.cmos.ca/congress2011/index.html - Special SPARC session

13-17 June 18th Conference on Atmospheric and Oceanic Fluid Dynamics, Spokane, Washington, USA,http://www.ametsoc.org/MEET/meetinfo.html

27 June - 8 July International Union of Geodesy and Geophysics Assembly “on” Earth on the Edge: Science for a Sustain-able Planet”. Melbourne, Australia, http://www.iugg2011.com

24-28 October WCRP Open Science Conference, Denver, CO, USA, www.wcrp-climate.org/conference2011

7-10 November 2011 NDACC Symposium, Reunion Island, France, http://www.reunion.fr

Design and Layout: D. Pendlebury and M. Rosen Editing: D. Pendlebury Printed and bound by: Thistle Printing - Canada ISSN 1245-4680

Edited and Produced by the SPARC IPOClimate-Chemistry Interactions: T. Peter (Switzerland),

A. R. Ravishankara (USA)

Stratosphere-Troposphere Dynamical Coupling: M. Baldwin (USA), S. Yoden (Japan)

Detection, Attribution, and Prediction of Stratospheric Change: W. Randel (USA), T.G. Shepherd (Canada)

Atmospheric Chemistry and Climate (AC&C): M. Chip-perfield (UK)

Gravity Waves: J. Alexander (USA)

Data Assimilation: S. Polavarapu (Canada)

CCM Validation Activity (CCMVal): V. Eyring (Ger-many), D. Waugh (USA), A. Gettelman (USA), S. Pawson (USA), T.G. Shepherd (Canada)

Laboratory Studies joint with IGAC: A.R. Ravishan-kara (USA), R. A. Cox (IGAC)

Solar Influences for SPARC (SOLARIS): K. Kodera (Japan), K. Matthes (Germany)

SPARC Dynamical Variability Activity: E. Manzini (Italy)

UTLS/SPARC Tropopause Initiative: P. H. Haynes (UK), A. Gettelman (USA), J. A. Añel (Spain)

The Role of Halogen Chemistry in Polar Stratospheric Ozone Depletion: M.J.Kurylo (USA), B.-M. Sinnhuber (Germany)

SPARC Data Initiative: M. Hegglin (Canada), S. Tegt-meier (Germany)

Themes and Activities

Future SPARC and SPARC-related Meetings

(January 11, 2011 / 08:57:26)

69967-1newsletter36-v2_New_p36.pdf .1

Figure 4: Monthly mean surface O3 (averaged between 12 h – 18 h local time) over North America for August 2006 as simulated by GEOS-Chem (a) without assimilation and (b) with the assimilation of TES observations. Tropospheric O3 profiles from TES were as-similated using a sequential suboptimal Kalman filter between 1 July and 31 August 2006. The assimilation increased surface O3 abun-dances across western North America, reflecting the influence of transport of background O3 (O3 not produced from North American emissions) from the free troposphere into the North American boundary layer. (From Parrington et al., 2009.)

Figure 5: Impact of top-down NOx emissions on surface O3 abundances in GEOS-Chem. (a) The original NOx emissions, in 1011 N cm-2

s-1, for August 2006. (b) Simulated surface O3 abundances based on the original NOx emissions. (c) The difference in the top-down and original NOx emissions. The top-down emissions were inferred from SCIAMACHY observations of NO2 (Martin et al., 2006) and sug-gest lower NOx emissions across the eastern USA. (d) The change in surface ozone abundances due to incorporating the top-down NOx emissions in GEOS-Chem. The reduced NOx emissions in the eastern USA produced a significant reduction in surface O3 abundances. (Figure courtesy of Mark Parrington.)

Report on the 7th SPARC Data Assimilation Workshop

Original NOx emissions Original surface O3

Change in NOx emissions Change in surface O3

55N

50N

45N

40N

35N

30N

25N 120W 100W 80W

55N

50N

45N

40N

35N

30N

25N 120W 100W 80W

55N

50N

45N

40N

35N

30N

25N 120W 100W 80W

55N

50N

45N

40N

35N

30N

25N 120W 100W 80W

0.00 0.70 1.40 2.10 2.80 3.50(1011 atoms cm-2 s-1)

-2.00 -1.00 0.00 1.00 2.00(1011 atoms cm-2 s-1)

0 10 20 30 40 50 60 70 80 90(ppbv)

-15 -10 -5 0 5 10 15(ppbv)

55N

50N

45N

40N

35N

30N

25N120W 100W 80W

60N

55N

50N

45N

40N

35N

30N

25N120W 100W 80W

60NGEOS-Chem before assimilation GEOS-Chem after assimilation

(ppbv)0 18 36 54 72 90

I

(January 11, 2011 / 08:55:43)

69967-1 SPARCNewsletter36-colour_New_p01.pdf .1

-15 -10 -5 0 5 10 15Lag (years)

-3

-2

-1

0

Sep

tem

ber s

ea-ic

e ex

tent

anom

aly

(mill

ion

km2 )

-1

0

1

2

3

ON

D T

air a

nom

aly

(°C

)

Tair trend

J F M A M J J A S O N D J

0

1

2

3

o C d

ecad

e-1

* * ** ** ** * ** ** ** *

Tair trend:during sea-ice loss periods

outside sea-ice loss periods

-3.0-2.0-1.0-0.50.00.51.02.03.0

oC decade-1

-3.0-2.0-1.0-0.50.00.51.02.03.0

oC decade-1

(a) (c)

(b)

during sea-ice loss (∆Taccel)outside sea-ice loss (∆Tavg)

120ºW

90ºW

60ºW

150ºW 180º 150ºE

120ºE

90ºE

60ºE

30ºE 0º 30ºE

Figure 6: Physical mechanisms of decadal pre-dictability associated with the production of Denmark Strait overflow water (DSOW), which is a major source of North Atlantic deep water. The North Atlantic inflows come through just two entry points, the Faroe-Shetland Chan-nel (FSC) and the Iceland-Faroe Ridge (IFR), and then are modified by surface fluxes while they transit through the Nordic seas. The Arc-tic Ocean and Barents Sea act as ‘switchyards’, adding decadal time scale delays to the system. These delays are variable in time and differ for surface and mid-depth waters. The latter feed the overflows and offer a predictive potential in the form of transient anomalies of the den-sity stratification. For the mid-depth, the figure shows a schematic circulation of Atlantic de-rived water (red solid) and dense, deep water (black dashed). From Karcher et al. (JGR, in revision).

Figure 7: Evidence for decadal-scale impact of sea-ice loss on Arctic land warming rates. (a) Composite anomaly time series of September sea-ice extent (solid line) and October-November-De-cember (OND) surface air temperature Tair (dashed line) over the Arctic land area (within 65–80°N, 60–300°E). Com-posites are formed by averaging nine 31-year anomaly time series that are cen-tred about the mid-point (lag 0 years) of a rapid sea-ice loss event simulated in a CCSM3 21st century A1B simulation. The individual time series are anoma-lies from the lag -10 to -5 year mean. (b) Average monthly Arctic land Tair trends during periods of rapid sea-ice loss compared to periods of moderate sea-ice loss. The asterisks indicate the months for which the differences in the trends are statistically significant at the 90% (single asterisk) and 95% (double aster-isk) levels. The largest impact is found in autumn and winter. (c),(d) Maps of Tair trends for OND during periods of rapid and moderate sea-ice loss. From Law-rence et al. (2008 GRL).

Report on WCRP Workshop on Seasonal to Multi-Decadal Predictability of Polar Climate

<

II

(January 12, 2011 / 13:29:27)

69967-1 SPARCNewsletter36-colour_New_p02rev2.pdf .1

Report on WCRP Workshop on Seasonal to Multi-Decadal Predictability of Polar Climate

Figure 1: An image of the PANSY radar.

Program of the Antarctic Syowa MST/IS Radar (PANSY)

1 Oct2008

1 Nov2008

1 Dec2008

1 Jan2009

1 Feb2009

1 Mar2009

1 Apr2009

20

30

40

50

60

70

80

Alti

tude

(km

)

NOx vmr 60 - 90N

1

1 5

5

5

5

5

5

5

55

2020

20

100 10

0

100

500

1000 K

2000 K

3000 K 4000 K

ppbv

0.1

1.0

10.0

100.0

1 Oct2008

1 Nov2008

1 Dec2008

1 Jan2009

1 Feb2009

1 Mar2009

1 Apr2009

20

30

40

50

60

70

80

Alti

tude

(km

)

CO vmr 60 - 90N

0.05

0.05

0.20

0.20

1.00

1.00

5.00

5.00

1000 K

2000 K

3000 K 4000 K

ppmv

0.1

1.0

10.0

Figure 2: Temporal evolution of NOx (top) and CO (bottom) as observed by MIPAS during the 2008/2009 NH winter at 70-90°N. The white dashed lines indicate selected potential tempera-ture levels.

The High-Energy-Particle Precipitation in the Atmosphere (HEPPA) Model vs. Data Inter-comparison:

Lessons Learned and Future Prospects

III

(January 11, 2011 / 08:55:49)

69967-1 SPARCNewsletter36-colour_New_p03.pdf .1

40N-45N 45S-40S(a) (b)

Alti

tude

(km

)

22

20

18

16

14

12

00:00:0001.01.06

00:00:0001.01.08

00:00:0001.01.07

Time (UTC)

Alti

tude

(km

)

22

20

18

16

14

12

00:00:0001.01.06

00:00:0001.01.08

00:00:0001.01.07

Time (UTC)

50302015109.08.07.06.05.04.03.02.01.00.0

SCIA H20, zonal average, vmr (ppmv)

Figure 6: Zonal mean water vapour values retrieved from SCIAMACHY limb measurements and ECMWF temperature and pressure averaged between (a) 40°N - 45°N and (b) 40°S -45°S. Every seventh day of SCIAMACHY measurements between August 2006 and August 2008 is used.

Ozo

ne H

ole

Are

a(m

illio

n sq

. km

)

40

30

20

10

01960 1980 2000 2020 2040 2060 2080 2100

Observations NIWAAMTRAC3CAM3.5CCSRNIESCMAMCNRM-ACME39CA

GEOSCCM

LMDZreproMRI

SOCOLULAQUMSLIMCATUMUKCA-METOUMUKCA-UCAMWACCM

EMAC-FUB

Niwa-SOCOLFigure 4: Simulated ozone hole areas based on the 1960-1965 minimum in the CCMVal pro-jections, including EMAC-FUB and E39C-A. Figure taken from Austin et al. (2010).

Stratospheric Change and its Role for Climate Prediction (SHARP): A contribution to SPARC

Water Vapour JA 360 KMIPAS

50N

40N

30N

20N

10N

0

10S20W 0 20E 40E 60E 80E 100E 120E 140E 160E 180

EMAC-FUB50N

40N

30N

20N

10N

0

10S20W 0 20E 40E 60E 80E 100E 120E 140E 160E 180

E39CA50N

40N

30N

20N

10N

0

10S20W 0 20E 40E 60E 80E 100E 120E 140E 160E 180

5 10 15 20 30 35 40ppmv

45 50 55 60 65 70 75 80 85 90 95 100256 8 10 12 14 16 18 20 22 24 26 28 30 32ppmv

Figure 5: Water vapour (ppmv) for July-August at 360 K for the region 10°S–50°N, 20°W–180°E. Left: 4 years of MIPAS data; overlaid as streamlines are the horizontal wind components of ECMWF analyses. Other panels: Long-term monthly mean water vapour (ppmv) (41/44 yr) for the two CCMs as indicated. Note the differing absolute values and colour scales in the MIPAS observational distributions and the CCM results, respectively. Figure updated from Kunze et al., 2010.

IV

(January 11, 2011 / 08:55:57)

69967-1 SPARCNewsletter36-colour_New_p04.pdf .1